Understanding multiscale structure–property correlations in PVDF-HFP electrospun fiber membranes by SAXS and WAXS

Electrospinning is a versatile technique to produce nanofibrous membranes with applications in filtration, biosensing, biomedical and tissue engineering. The structural and therefore physical properties of electrospun fibers can be finely tuned by changing the electrospinning parameters. The large parameter window makes it challenging to optimize the properties of fibers for a specific application. Therefore, a fundamental understanding of the multiscale structure of fibers and its correlation with their macroscopic behaviors is required for the design and production of systems with dedicated applications. In this study, we demonstrate that the properties of poly(vinylidene fluoride-co-hexafluoro propylene) (PVDF-HFP) electrospun fibers can be tuned by changing the rotating drum speed used as a collector during electrospinning. Indeed, with the help of multiscale characterization techniques such as scanning electron microscopy (SEM), small-angle X-ray scattering (SAXS), and wide-angle X-ray scattering (WAXS), we observe that increasing the rotating drum speed not only aligns the fibers but also induces polymeric chain rearrangements at the molecular scale. Such changes result in enhanced mechanical properties and an increase of the piezoelectric β-phase of the PVDF-HFP fiber membranes. We detect nanostructural deformation behaviors when the aligned fibrous membrane is uniaxially stretched along the fiber alignment direction, while an increase in the alignment of the fibers is observed for randomly aligned samples. This was analyzed by performing in situ SAXS measurements coupled with uniaxial tensile loading of the fibrous membranes along the fiber alignment direction. The present study shows that fibrous membranes can be produced with varying degrees of fiber orientation, piezoelectric β-phase content, and mechanical properties by controlling the speed of the rotating drum collector during the fiber production. Such aligned fiber membranes have potential applications for neural or musculoskeletal tissue engineering.


Introduction
Electrospinning is a simple, and cost-effective method to produce ber membranes for various applications such as ltration, drug delivery, biosensing, and tissue engineering due to the inherent advantages of a high surface-to-volume ratio and high porosity. [1][2][3][4][5][6][7][8] Fiber membranes have gained ample attention of researchers in recent times as a scaffold for tissue engineering applications due to their abilities to mimic the extracellular matrix of natural tissues more accurately than macro-or micro-scale biomaterials. [9][10][11][12] For example, aligned ber membranes have shown promising response for neural and musculoskeletal tissue engineering as aligned bers signicantly induce neurite outgrowth and promote cell migration compared to randomly oriented ber membranes. [13][14][15][16][17][18][19] However, scaffolds for neural and musculoskeletal tissue engineering require the optimization of the degree of ber alignment, a piezoelectric response, and good mechanical properties. 13,14,17,20,21 It is well known that the various parameters of the electrospinning process such as the applied potential difference, the type of sample collector, the needle diameter, the needle to collector distance, the ow rate and the environmental conditions critically inuence ber properties. 11,[22][23][24] This brings various possibilities to steer and tailor electrospun ber membranes for the required application. 11 At the same time, such a large parameter window and their nonlinear dependance on the morphology and properties of the bers makes electrospinning a challenging process for obtaining the desired properties. 25,26 Therefore, the understanding of the multiscale structure-property relationship is essential to steer the properties of electrospun brous membranes accurately.
In the past, different techniques have been developed by applying external conditions such as mechanical, electrical, and magnetic forces to align bers in a desired direction. [27][28][29] One of the most straight forward ways to align bers is to use a rotating drum collector. 28,30,31 Several studies are reported to use a rotating drum collector to produce ber membranes with varying degrees of orientation of the bers within the membrane. 28,32,33 However, producing electrospun bers with desired properties (e.g., degree of orientation, mechanical, surface properties, etc.) is primarily based on the trial-and-error method. In a previous study, we combined scanning electron microscopy (SEM), atomic force microscopy (AFM), small-angle X-ray scattering (SAXS) and wide-angle X-ray diffraction (WAXD) to decode the multiscale structure of non-aligned and aligned poly(vinylidene uoride-co-hexauoro propylene) (PVDF-HFP) electrospun ber membranes at various speeds of the rotating drum. 30 We reported that electrospun bers possess a nano-brillar structure with repeating lamellae with tie molecules in between and concluded that the moderate rotating speed of the rotating drum collector used in the study only inuences the alignment of the bers in the membranes but results in no change in the internal structure of the bers. 30 However, Kongkhlang et al. have reported that electrospinning can be used for controlling the morphology and crystal orientation in polyoxymethylene (POM) bers by producing them at different rotating collector speeds. 34 They reported that increasing the rotation speed of the rotating drum collector results in additional stretching on the bers. 34 This inuences the molecular arrangement of the polymeric chains within the bers.
Therefore, in this follow-up study, we hypothesized to produce electrospun bers at a rotating drum collector speed 2.2 times higher (23.0 m s À1 ) than in our previous study, 30 and the structural changes were compared with a lower rotating drum collector speed of 0.5 m s À1 . We expected that the additional stretching due to an extremely high rotating speed of the drum collector will induce more piezoelectric b-phase of PVDF-HFP due to change in molecular conrmation as observed in the stress-induced deformed PVDF lms and bers. [35][36][37] Furthermore, it has also been reported that the degree of alignment of the bers in the membranes provides enhanced mechanical properties along the ber alignment direction. 38,39 Hence, a high rotating speed of the drum collector might result in the change of the mechanical properties and the deformation behavior of ber membranes due to structural modications. Consequently, in this study, we present the effect of the highspeed rotating drum on the mm-nm-Å scale structural changes and their correlation with the mechanical and piezoelectric properties of the PVDF-HFP membranes. SEM, SAXS and WAXS were used to quantify multiscale structural changes and b-phase content in both membranes. The mechanical properties and mechanical deformation behavior of the two membranes produced at a low and an extremely high rotating speed of the collector were investigated by in situ SAXS measurement using uniaxial tensile loading up to 100% strain.

Materials and methods
Electrospun ber fabrication PVDF-HFP polymer (M w $400 000 g mol À1 ) and dimethylformamide (DMF) were purchased from Sigma Aldrich, Switzerland. PVDF-HFP pellets were dissolved in DMF resulting in a concentration of 35 w/v%. The mixture was stirred overnight on an orbital shaker to prepare a homogeneous polymer solution. A 3 ml syringe was used in the electrospinning setup with a blunt 21 G needle. A rotating drum collector of 10 cm diameter was used with 50 and 2200 rpm that correspond to 0.5 and 23.0 m s À1 linear speeds, respectively. The homogeneous polymer solution was emitted at a rate of 20 ml min À1 under a potential difference of 19 kV (+14/À5 kV) from a needle-tocollector distance of 25 cm. Electrospinning was performed at room temperature ($23 C) and with about 55% humidity.

Scanning electron microscopy (SEM)
A scanning electron microscope (Hitachi S-4800, Hitachi-High Technologies, Illinois, USA) with an acceleration voltage of 5 kV and a current ow of 10 mA was used to image the ber membranes. An 8 nm think layer of conductive gold-palladium was deposited on each sample before imaging. DiameterJ (a plugin for imageJ) was used to determine the diameter of the bers. 40 The average diameter of the bers was determined by taking an average of 30 bers from SEM micrographs.
Small-angle X-rays scattering (SAXS) and wide-angle X-ray Scattering (WAXS) SAXS and WAXS measurements were performed with a Bruker Nanostar instrument (Bruker AXS GmbH, Karlsruhe, Germany). This pinhole-collimation system was equipped with a microfocused X-ray Cu source (wavelength Cu Ka ¼ 1.5406Å) and custom-built semitransparent beamstop providing a beam size of about 400 mm. A 2D MikroGap technology-based detector (VÅNTEC-2000) with 2048 Â 2048 pixels with a pixel size of 68 Â 68 mm 2 was used. The SAXS experiment was performed at a sample-to-detector-distance (SDD) of 107 cm providing a resolvable scattering vector modulus (q) range from 0.07 to 2.20 nm À1 . The WAXS experiment was performed at a SDD of 5 cm providing a resolvable q-range from 2 to 27 nm À1 . The SDDs were calibrated with silver behenate and corundum for the 107 cm and 5 cm distances, respectively. Each sample and background (residual air) were measured for 2 hours for both SAXS and WAXS under moderate vacuum conditions (10 À2 mbar pressure) to reduce air scattering. Aer the measurement, 1D proles were extracted using DIFFRAC.EVA (Bruker AXS, version 4.1).
In situ SAXS experiment: uniaxial tensile loading of the electrospun membranes A tensile stage (Anton Paar, TS600) was used for in situ SAXS investigations of the structural deformation in electrospun ber membranes under uniaxial tensile loading. A load cell capable of applying a tensile force in the range of 0 to 6 N was used in this experiment. A rectangular-shaped sample of 23.6 Â 10 mm was mounted on the tensile stage. Aerward, the tensile stage was placed inside the Bruker Nanostar instrument (Bruker AXS GmbH, Karlsruhe, Germany) to perform in situ SAXS measurements. The tensile stage setup inside the sample chamber of the Nanostar instrument is shown in Fig. 4(a). The ber membranes were stretched along the ber alignment direction. The in situ SAXS experiment was performed at a SDD of 107 cm. The mounted electrospun membranes were continuously elongated with a waiting time of 30 minutes at 2.5, 5, 10, 25, 50, 75 and 100% elongation when SAXS frames were recorded. 1D-proles were extracted using DIFFRAC.EVA (Bruker AXS, version 4.1).

Orientation analysis of bers inside the membranes
A new spatial cross-correlation method-based algorithm was further extended from our previous study to quantify the degree of alignment of bers in the electrospun membranes from SEM images. 30 Small dimensions (500 Â 500 pixels) were taken from the SEM images at a random position and were used as reference images. The spatial cross-correlation coefficients (R) were calculated with cropped images of same dimensions by shiing pixel-by-pixel in each direction from À90 to +90 with 1 step size. Thereaer, the number of pixel values at half decay of the correlation coefficient (N 1/2 ) was determined by exponential tting for each rotating direction. To improve the statistics, the same procedure was repeated by randomly selecting 50 cropped boxes on the SEM images. The average of N 1/2 was determined for each rotating direction; the N 1/2 vs. the orientation angle is obtained. Finally, the resulting curve was tted with a Lorentzian function and the respective full width at half maximum (FWHM) was determined. The obtained FWHM was correlated with the misorientation width (MOW) of the bers inside the membranes.
Porod-correlation peak model tting of SAXS data A Porod-correlation peak model was used to t the 1D-SAXS proles aer the background subtraction. The Porodcorrelation peak model is given in eqn (1) and the details are discussed elsewhere: 30 where A and B are the multiplying factors, a is the Porod exponent, q 0 is the peak position, u is the FWHM, and m is the tting parameter for the peak shape correction. All ttings were performed in Matlab (R2019b). The average lamellar spacing (d) was determined by using eqn (2). Fitting of WAXS data to quantify the a/b-phase ratio PVDF is co-polymerized with hexauoropropylene (HFP), but due to the low crystallinity of the PVDF-HFP co-polymer, no crystal structure is available in the crystallographic database. However, PVDF and PVDF-HFP have been considered to possess similar crystal structures. 41,42 Therefore, the monoclinic PVDF a (P2 1 /c) and b (Cm2m) phase crystallographic information provided by the Cambridge Crystallographic Data Center (CCDC) no. 1207416 and 1207418, respectively, was used for tting. 43 A total of 5 peaks (namely, amorphous, (100/020) a , (110) a , (100/200) b and (021/111/120) a ) were selected and tted with a Gaussian function. Guan et al. also reported on using these main peaks to quantify the a and b-phases in the PVDF-HFP lms. 42 Due to the approximations used in our tting, we aim only to highlight the relative changes between the patterns, in particular towards the a/b phase ratio, and do not report on absolute values.

Results and discussion
Microscale morphology of the ber membranes The parameters for the electrospinning of PVDF-HFP were optimized in our previous study. 30 In this study, the rotating drum was set at a speed of 0.5 m s À1 and 23.0 m s À1 leading to the production of homogeneous non-beaded ber membranes.
The SEM micrographs of the ber membranes are shown in Fig. 1(a) and (b) and macroscopic image in ESI Fig. S1. † The bers obtained with the low rotating speed led to a randomly oriented ber mesh, while with the high rotating speed the bers were preferentially oriented. The bers within the membrane electrospun at 0.5 m s À1 exhibited a diameter of 1.6 AE 0.1 mm, whereas the bers electrospun at 23.0 m s À1 exhibited a diameter of 1.1 AE 0.1 mm. It means that the diameter of the bers decreased at a sufficiently high rotation speed of the collector. It is, in general, proposed that the strong stretch force induced due to the high-speed rotating drum is responsible for the decrease in the diameter of bers. 32,39 Such a strong force might induce structural changes within the bers as observed by Kongkhlang et al. 34 Determination of the degree of orientation using SEM micrographs of the membranes The degree of orientation of the bers inside the membranes was determined based upon our new spatial cross-correlation method-based algorithm which is shown in Fig. 1(c) and (d) for À90 to 90 orientation angles for the membranes produced at 0.5 m s À1 and 23.0 m s À1 . The orientation angle of 0 is considered as the reference for the ber alignment direction.
The membrane produced at 0.5 m s À1 does not show any peak due to the random orientation of the bers (Fig. 1(c)). A peak about 0 is observed for the membrane produced at 23.0 m s À1 rotation speed due to the orientation of the bers (Fig. 1(d)); a MOW of (22.2 AE 2.8) is obtained by peak tting.

Nanostructure modication observed by SAXS
The 2D-SAXS proles of the electrospun ber membranes at 0.5 m s À1 and 23.0 m s À1 are shown in Fig. 2(a) and (b) respectively. An isotropic scattering prole is observed for the sample produced at 0.5 m s À1 speed due to the random orientation of the bers inside the membranes. However, the 2D-SAXS prole for the electrospun ber membrane produced at 23.0 m s À1 speed depicts an anisotropic scattering prole related to the alignment of the bers. The peak observed in the 2D-SAXS prole along the ber alignment direction reect the periodicity of the crystalline (lamella) and amorphous domains within the bers for the sample produced at 23.0 m s À1 . Similarly, a halo is observed for the sample produced at 0.5 m s À1 due to the random distribution of bers within the membranes. The center position of the peak or halo corresponds to the average distance, also known as lamellar spacing, between two adjacent lamella domains. In order to quantify the change in the lamellar spacing, 1D-radial proles were extracted by performing azimuthal integration (30 wedge as indicated in Fig. 2(a) and (b)) on the 2D-SAXS patterns along the ber alignment direction. The corresponding 1D extracted radial prole tted with the Porod-correlation peak model given by eqn (1). The lamellar spacing is shown in Fig. 2(c) and (d). The peak position provides the lamellar spacing corresponding to the average distance between two adjacent lamella domains. The periodic repetition of lamella  44 where the atom colours grey, yellow, and white represent C, F, and H, respectively. domains constitutes the nanobrillar structure inside the electrospun bers. 30 A decrease in the lamellar spacing is observed for the bers electrospun at an extremely high speed as shown in Fig. 2(d). However, no change in the lamellar spacing was observed in our previous study when bers were produced at moderate rotating drum speeds up to 10.5 m s À1 . 30 This reects that the polymeric chains of PVDF-HFP in the nanobrillar structure of the bers tend to be more aligned related to further mechanical stretching of the bers collected at an extremely high rotating speed while negligible mechanical stretching occurs when the rotating speed was moderate. Such behavior of the rotating drum speed is due to the nonlinear dependence of the electrospinning parameter on the ber morphology. 25,26 Moreover, the decrease in lamellar spacing (increased polymer chain stretching) is the reason for the decrease in the diameter of the bers.
Quantication of the a/b-phase ratio using WAXS data WAXS measurements were performed to gain information about the relative amount of the aand b-phases in the samples. The 2D-WAXS proles of ber membranes produced at 0.5 m s À1 and 23.0 m s À1 rotating drum collector speeds are shown in Fig. 3(a) and (b) respectively. Similar to SAXS, isotropic diffraction rings were observed for the sample produced at 0.5 m s À1 speed due to the random orientation of the bers inside the membranes, while the anisotropic diffraction pattern for the ber membrane produced at 23.0 m s À1 is related to the preferential alignment of the bers.
In order to characterize the evolution of the aand b-phase fractions of the two membranes, the 1D-radial proles were extracted laterally with respect to the ber alignment direction from the 2D-WAXS data, which is shown in Fig. 3(c) and (d). Molecular drawings of aand b-phases of PVDF including the respective unit cell information are shown in Fig. 3(e). A tting was performed on the extracted 1D-radial proles, which are shown in Fig. 3(c) and (d). The broad amorphous peak represents the amorphous phase centered at 18.8 for both samples.
The peaks (100/020) a , (110) a and (021/111/120) a are at 17.8 , 19.8 , and 26.3 and represent the convoluted (100/020), (110) and (021/111/120) diffraction peaks of the a-phase, respectively. The peak (110/200) b is centered at 20.3 , which represents the convoluted (110/200) diffraction peak of the b-phase. The peak positions correspond to the unit cell of aand b-phases of PVDF as shown in ESI Fig. S2. † The relative fraction of aand b-phases was determined for both samples from the ratios of the area under peaks. The area under peaks and related determined phases are summarized in supported Table S1. † We obtained an increase of about 8 times for the piezoelectric b-phase for the sample produced at 23.0 m s À1 compared to 0.5 m s À1 ber membranes. The increase in the b-phase is because of the change in the molecular conformation (all-trans) induced due to the additional stretching of the polymer chains at an extremely high rotating speed. 35 To support the increase of the b-phase in the bers produced at 23.0 m s À1 compared to 0.5 m s À1 , Fourier-transform infrared spectroscopy (FTIR) measurements were performed. An increase in the b-phase in the FTIR plot could be also observed for 23.0 m s À1 with respect to the intensity of the corresponding peaks in the recorded spectra (see ESI Fig. S3 †). Even though FTIR spectra showed distinct band a-phase absorption bands, we relied on WAXS to determine the b-phase contents in the PVDF-HFP bers due to the unavailability of the absolute absorption coefficients for PVDF-HFP required for calculation.

Mechanical properties and deformation behavior of the membranes
In situ SAXS coupled with tensile loading. To further understand and explain the changes in mechanical properties and related electrospun membrane deformation behaviors with respect to the uniaxial tensile loading, in situ SAXS measurements were performed. The tensile loading was applied along the ber alignment direction. The tensile stage setup with the loaded electrospun membrane inside the sample chamber of SAXS instrument is shown in Fig. 4(a). The engineering stress- strain curves are shown in Fig. 4(b) and (c) for bers produced at 0.5 m s À1 and 23.0 m s À1 , respectively. Some relaxation (decrease in stress under constant strain) in the engineering stress-strain curve is observed due to the waiting time, while the SAXS experiment was performed for 30 minutes. The engineering stress-strain curve represents elastic as well as plastic domains for both ber membranes. Similar stress-strain behavior was also observed by Yano et al. for poly(vinyl alcohol) (PVA) 45 and Camarena-maese et al. for polycaprolactone (PCL). 46 The Young's modulus, yield stress, and yield strain were determined from the engineering stress-strain curve. The values are given in Table 1.
An increase in Young's modulus and yield stress was obtained for the ber membrane produced at 23.0 m s À1 compared to the 0.5 m s À1 rotating drum collector speed because aligned bers provide much higher resistance to the applied stress along their alignment direction. 34,46 Interestingly, the yield strain remains the same for both membranes.
The 2D-SAXS proles corresponding to an increasing percentage of engineering strain for electrospun membranes produced at 0.5 m s À1 and 23.0 m s À1 speeds are shown in Fig. 5(a) and (b) respectively. For the membrane produced at 0.5 m s À1 speed, the isotropic scattering gradually changes to anisotropic scattering upon increasing uniaxial tensile loading  ( Fig. 5(a)). It means that an increased number of bers start to align along the loading direction. 45,46 At 100% engineering strain, lamellar peaks start to appear more clearly in the 2D-SAXS prole. However, the membrane produced at 23.0 m s À1 speed was already aligned; therefore, when a uniaxial loading was applied, a streak-like signal started to increase while the peaks start to move towards the small q direction and nally disappear when 100% engineering strain is achieved (Fig. 5(b)). It could also be noticed that the intensity of the peaks increases up to 10% strain and then starts to decrease on further straining ( Fig. 5(b)). This is direct evidence of a modication in the polymer chain arrangement on the nanoscale within the bers.
To quantify the changes in the lamellar spacing on increasing engineering strain, the 1D-radial proles were extracted in the ber alignment direction by performing 60 wedge azimuthal integrations. The extracted 1D-radial proles were tted with the same Porod-correlation peak model given by eqn (1). The 1D extracted SAXS radial prole together with the tting is shown in Fig. 6(a) and (b) for the respective ber membranes. The determined lamellar spacings are given in Fig. 6(c).
The lamellar spacing for the membranes produced at 0.5 m s À1 speed remains constant in the elastic region, while a slight increase with a plateau is observed in the plastic deformation domain. Fibers inside the membranes produced at 0.5 m s À1 speed under a uniaxial tensile loading up to 100% of engineering strain mainly get aligned with only little internal nanostructure modication in the plastic domain. However, the lamellar spacing for the membranes produced at 23.0 m s À1 speed increases linearly before reaching a plateau, followed by a decrease at 100% strain. In the case when most of the bers were aligned, the uniaxial tensile loading along the alignment direction of the bers is able to modify their nanostructure. The increase in the lamellar spacing is due to the elongation of the polymer chains in the amorphous region between the lamella domains. 47 However, when the amorphous region with entangled polymer chains reaches the maximum elongation, then further increase in lamellar spacing is not possible and further stretching leads to the fracture of bers. Consequently, lamellar spacing reaches a plateau and the intensity of the correlation peak also decreases. 46 The decrease in lamellar spacing and the intensity of the correlation peak at 100% strain is probably due to the loss of electronic contrast between the lamella and amorphous domains as most of the polymer chains from amorphous domains get aligned due to stretching.
SEM micrographs of ber membranes aer 100% elongation in the relaxed state. SEM images were taken for both membranes aer 100% uniaxial tensile loading in the relaxed state to observe the nal stage and to compare with the SEM images of the pristine ber membranes ( Fig. 1(a) and (b)). The SEM images of the 100% strained membranes produced at 0.5 m s À1 and 23.0 m s À1 speeds are shown in Fig. 7(a) and (b) respectively, while Fig. 7(c) and (d) show the corresponding zoomed-in regions showing the ber-to-ber connection marks. The ber membranes produced at 0.5 m s À1 speed still show entangled ber networks even aer 100% strain, while bers produced at 23.0 m s À1 speed are highly aligned. Oval-shaped marks were observed for both ber membranes (Fig. 7(c) and (d)) meaning that ber-to-ber connections could only be physical and no chemical bonds were formed during electrospinning. 48 Therefore, when a tensile force was applied, the bers were able to move apart leaving these oval-shaped marks behind. These marks also indicate poor evaporation of the DMF solvent during electrospinning.
Moreover, no fractured bers were observed for both membranes up to 100% strain meaning that more force is required to achieve a fracture of the bers. However, the bers of the stretched membranes produced at 0.5 m s À1 have an average diameter of 1.5 AE 0.1 mm and the bers of the stretched membrane produced at 23.0 m s À1 have an average diameter of 0.9 AE 0.1 mm. In comparison with the non-stretched bers, the average diameter for the sample produced at 0.5 m s À1 remains unchanged, while a slight decrease was observed for the bers produced at 23.0 m s À1 . This can be explained by the fact that the bers produced at 23.0 m s À1 were further stretched upon uniaxial tensile loading, which resulted in nanostructure modication as observed in the in situ SAXS experiment.

Conclusion
This study shows that the ber diameter, degree of alignment, b-phase content, and consequently the mechanical properties of electrospun PVDF-HFP ber membranes can be steered by changing the speed of the rotating drum collector. The bers produced at a higher rotating drum speed (23.0 m s À1 ) possess smaller diameters, a higher degree of alignment, and a change in polymer chain arrangement, which induces more piezoelectric b-phase content compared to bers produced at a low speed (0.5 m s À1 ). The degree of orientation of bers was quantied from SEM micrographs by the extension of our new image crosscorrelation method-based algorithm. The increase in lamellar spacing was quantied by SAXS for bers produced at 23.0 m s À1 and compared to 0.5 m s À1 . An increase of about 8 times of the b-phase was obtained by tting the WAXS data for membranes produced at 23.0 m s À1 compared to 0.5 m s À1 . Moreover, the membranes electrospun at 0.5 m s À1 became more aligned aer stretching but the diameter of the bers did not change even at an elongation up to 100%. In addition, the membrane electrospun at 23.0 m s À1 demonstrated a change in the nanostructure of the bers according to the in situ SAXS analysis. The diameter of the bers decreased for this membrane at 100% elongation compared to the pristine ber sample. This study contributes to the fundamental understanding of the multiscale structure-property relationship for producing electrospun ber membranes with steered properties. The aligned ber membranes investigated in this study with enhanced piezoelectric b-phase content and mechanical properties have potential applications in the eld of neural or musculoskeletal tissue engineering.

Conflicts of interest
There are no conicts to declare.