C. Merliniab,
G. M. O. Barra*a,
T. Medeiros Araujob and
A. Pegoretti*b
aMechanical Engineering Department, Federal University of Santa Catarina (UFSC), Florianopolis, SC, Brazil. E-mail: g.barra@ufsc.br
bDepartment of Industrial Engineering, University of Trento, 38123 TN, Italy. E-mail: alessandro.pegoretti@unitn.it
First published on 17th March 2014
Non-woven mats with highly pressure-sensitive electrical conductivity have been prepared by electrospinning of poly(vinylidene fluoride) (PVDF) containing up to 23 wt% of polypyrrole (PPy) particles synthesized by using dodecylbenzenesulfonic acid (DBSA) as a dopant. The obtained mats have been characterized by dynamic mechanical thermal analysis, differential scanning calorimetry, thermogravimetric analysis, infrared spectroscopy, scanning electron microscopy and nitrogen adsorption BET. Electrical resistivity changes of PVDF/PPy blends were investigated during loading–unloading compressive cycles. It has been observed that the electrical resistivity varies reversibly with the applied compressive stress. The maximum sensitivity was obtained for a PVDF/PPy blend containing 13 wt% of PPy, manifesting an electrical resistivity drop of 10 orders of magnitude, i.e. from 1017 to 107 Ω cm upon application of a compressive stress of 5 MPa. This peculiar response can be attributed to the formation of stress-induced conducting pathways in the electrospun network. Considering the remarkable resistivity change and the reproducibility of the phenomenon after repeated loading–unloading cycles, this mat may find application as a pressure sensor.
The effect of a compressive stress on the resistivity of conductive polymeric composites produced from bulk material has been studied by several authors.1,2,8,10,11 In most cases, electrical resistivity variations between 1 and 5 orders of magnitude have been reported, usually with resistivity values ranging between 103 and 108 Ω cm. Additionally, the majority of these pressure sensitive materials display some degree of hysteresis after subsequently compression cycles.1,2,11 Several strategies have been investigated to improve the sensitivity and reproducibility of these conductive polymeric systems. A possible alternative is the development of polymeric membranes composed of fibers coated by intrinsically conducting polymers (ICP's) such as polypyrrole (PPy) and polyaniline (PANI).12,13 These materials can be commonly prepared using in situ oxidative processes or electrochemical methods in the presence of natural or synthetic insulating polymer fibers. Electromechanical tests showed that ICP's coated fibers can be used as pressure-sensitive devices, combining the electrical conductivity with good compression sensitivity for low amounts of conducting polymer.12
Another promising alternative is the development of pressure sensing devices using nano-structured materials.14 The electrospinning technique can be used to produce polymeric nano-fiber mats, leading to the possibility of obtaining new nanostructured composites with an improved sensitivity and faster response. In fact, advantages can be offered by their very large surface area to volume ratio (up to 103 times larger in comparison to micro-fibers),15 high density of sensitive sites and porous structure, associated with the properties of conductive polymers.16–19 In the electrospinning process, an elevated electric field (of about 1 kV cm−1) is applied to a needle through which a polymer solution is delivered. When the applied electric field overcomes the surface tension and the viscoelastic forces in the droplet, a charged jet of the polymer solution is ejected. While traveling to the collector, the jet exhibits bending instabilities, the solvent evaporates and solidified fibers are deposited on a collector.15,20 Conductive sub-micron fiber mats can be produced by (i) direct electrospinning technique, in which fibers are electrospun from a mixture containing an insulating polymer and ICP dispersed in a common solvent17,19,21–23 or by (ii) an indirect method, in which the insulating fibrous membrane is firstly prepared by electrospinning and after coated with PPy or PANI through pyrrole or aniline oxidative polymerization.18,24–26
In this context, poly(vinylidene fluoride) (PVDF) is a suitable polymer matrix to prepare electrospun conductive polymer composites due to its unique pyroelectric/piezoelectric features, coupled with superior mechanical properties and easy processability.16 Highly sensitive piezoelectric sensors based on PVDF films and nanofibers have been successfully obtained and described by several authors.27–31 These materials are commonly used in piezoelectric devices in order to measure forces and strains, at a micron-scale in terms of displacement, especially those of dynamic nature, such as vibrations, accelerations and oscillations.14 Such devices present high sensitivity in the low-pressure regime (<10 kPa). However, to the best of our knowledge, there are no studies published in the open scientific literature concerning the development of electrospun conductive mats based on PVDF/PPy for pressure sensors applications. The PVDF/PPy mats developed in the present study can be employed in a sensor capable to generate an outstanding variation of the electrical resistivity at high-pressure levels (0.1–5 MPa). Unlike the PVDF sensors, that are based on the piezoelectric concept, the PPy particles inside of the PVDF fibers may form a conducting network under pressure, according with a percolative theory. Additionally, the use of these materials with sensitivity at high pressure avoid signal shift due to undesirable noise and/or interference, such as mechanical vibrations or thermal expansion.27
Therefore, the focus of this study is on the preparation of conductive sub-micron PVDF/PPy fiber mats through electrospinning technique and on the investigation of their microstructure, thermomechanical properties and stress-dependent conductivity under cyclic loads.
The specific surface area (SSA) of the electrospun samples was determined by nitrogen adsorption BET method (ASAP 2010, Micromeritics, USA).
The viscosity of the solutions was evaluated at 25 °C using a Brookfield viscometer Model RVT 90658 using SC4-21 spindles. The shear rate was fixed at 0.46 s−1.
Attenuated-total-reflectance Fourier-transform-infrared (ATR-FTIR) spectroscopy was performed on a Bruker spectrometer, model TENSOR 27, in the range of wavenumbers 4000–400 cm−1 by accumulating 32 scans at a resolution of 4 cm−1.
The microstructure of electrospun mats of neat PVDF and PVDF/PPy blends was analyzed by a Zeiss-Supra 60 field emission scanning electron microscope (FESEM), and by a Jeol JSM-6390LV scanning electron microscope (SEM). The samples were sputtered with gold and observed using a secondary electron detector and applied tension between 2 and 4 kV.
Thermogravimetric analysis (TGA) was carried out using a TGA Q5000IR (TA Instruments, USA) thermo-gravimetric analyzer. The analyses were performed at 10 °C min−1 from 35 °C to 700 °C under a nitrogen flux of 25 mL min−1.
Differential scanning calorimetry (DSC) measurements were performed on a Mettler DSC30-TA Low Temperature and a Mettler TC 15 TA Controller at a heating rate of 10 °C min−1, from −75 °C to 220 °C, under a nitrogen flux of 100 mL min−1. The crystallinity content (Xc) of PVDF was calculated on the basis of eqn (1).
(1) |
Dynamic mechanical thermal analysis (DMTA) experiments were carried on a DMA Q800 (TA Instruments) under the tensile mode on rectangular specimens with a width of 6.4 mm and a length of 30 mm. The analyses were performed at a frequency of 1 Hz, from −100 to 130 °C, at a heating rate of 3 °C min−1 and a peak-to-peak displacement of 64 μm.
The transmission electron microscope (TEM) experiments were performed on a JEM-1011 microscope with an acceleration voltage of 100 kV.
The setup used to evaluate the electrical–mechanical response of the PVDF/PPy blends consists in an universal testing machine (Instron, Model 5969) to apply a controlled stress and an electrometer (Keithley 6517A) to acquire the resistivity data, using a software developed exclusively for this work. The interface RJ45 and GPIB were used to connect the Instron and electrometer to the computer, respectively. The specimen was placed between two electrodes (Ø = 22.5 mm) which were confined in a cylinder made of poly(tetrafluoroethylene) in order to electrically isolate the sample from the mechanical testing fixtures. The device was then placed between the testing plates of the universal testing machine. The electrodes were connected to the electrometer for measuring the volume resistivity of the blends during loading and unloading. The samples were loaded up to 5 MPa at a loading rate of 1 MPa min−1, the compressive stress was then released at the same rate. For each sample, loading–unloading sequences of 5 and 25 cycles were performed on different specimens. The electrical resistivity (ρ) value (in Ω cm) were calculated through eqn (2), where R is the measured resistance (Ω), d is the sample diameter (cm) and w is the specimen thickness (cm).
(2) |
Nominal PPy content in the blend (wt%) | Measured PPy content in the blenda (wt%) | Apparent viscosity (mPa s) | Electrical resistivity × 1014 (Ω cm) | Specific surface area (m2 g−1) | Fiber diameter (nm) |
---|---|---|---|---|---|
a From CHN elemental analysis. | |||||
0 | 0 | 2476 | 27.00 ± 1.20 | 7.59 ± 0.05 | 854 ± 271 |
3 | 5.11 ± 0.07 | 2485 | 24.10 ± 3.40 | 6.65 ± 0.08 | 968 ± 280 |
13 | 13.61 ± 0.22 | 6520 | 1.39 ± 1.31 | 15.88 ± 0.39 | 790 ± 170 |
23 | 26.90 ± 0.44 | 48830 | 1.60 ± 1.27 | 22.16 ± 0.33 | 460 ± 90 |
Fig. 1a–d illustrates FESEM images of electrospun mats of neat PVDF and PVDF/PPy blends containing various amounts of PPy. Micrographs show that both PVDF (Fig. 1a) and PVDF/PPy mats (Fig. 1b and c) are constituted by a three-dimensional fibrous network. Neat PVDF fiber mats displays an average diameter of 854 ± 271 nm. It is worthwhile to note that the average diameter of PVDF/PPy electrospun mats decreases as the PPy content in the blend increases, with an average value of 460 ± 90 nm for the PVDF/PPy_23 specimen. The addition of PPy induces a modification of the properties of the solutions, such as viscosity and electrical conductivity. As a consequence, the voltage applied during the electrospinning process has to be adjusted, from 15 to 23 kV with the increasing of PPy content. One of the effects of increasing the voltage is the reduction of the fiber diameter which can be observed as the amount of PPy increases. As the amount of PPy is increased from 3 to 13 wt% the formation of PPy agglomerates along the electrospun fibers can be noticed, even if no presence of PPy particles outside the electrospun fibers can be observed. A non-woven structure with the existence of larger agglomerates along the fibers is formed in the blends containing 23 wt% of PPy. Nevertheless, the micrographs of the last one revealed an incidence of a consistent quantity of PVDF/PPy agglomerates outside of the non-woven structure. The importance of this peculiar microstructure to reach a stress-induced electrical conductivity will be further discussed.
Fig. 1 FESEM micrographs of electrospun mats of (a) neat PVDF, (b) PVDF/PPy_3, (c) PVDF/PPy_13 and (d) PVDF/PPy_23. |
FTIR spectra of PPy, PVDF and PVDF/PPy fiber mats are reported in Fig. 2. The spectrum of PPy exhibits absorption bands at 1520 cm−1 and 1424 cm−1, assigned to C–C and C–N stretching vibration of pyrrole ring, respectively.11,35 The absorption band at 1275 cm−1 is assigned to C–H or C–N in-plane deformation modes,36 while the bands at 1121 and 1015 cm−1 are assigned to the C–H bending modes. The band centered at 960 cm−1 is related to the C–H out-of-plane deformation of pyrrole ring.11,35 An interesting feature of PVDF is its polymorphism, i.e. the possibility to form different crystalline structures, such as, alpha (α), beta (β), gamma (γ) and delta (δ), depending on the crystallization conditions.37 The various PVDF phases can be identified by the absorption bands in the FTIR spectrum. Electrospun PVDF spectrum shows absorption bands at 1402 and 876 cm−1 (band for amorphous phase) attributed to the C–F stretching vibration and at 1172 cm−1 assigned to the C–C bond.24,37–39 The bands at 837, 1073 and 1274 cm−1 are associated with the β phase.34,39–41 The bands at 603 and 1232 cm−1, are characteristics of α and γ phases, respectively.39 The FTIR spectra of PVDF/PPy blends show absorption bands similar to those of neat PVDF, however, the blend with 23 wt% of PPy presents a stronger band at 1540 cm−1. This behavior can be related to the presence of PVDF/PPy agglomerates outside of the fibers, as observed in the FESEM micrographs. The electrospun fibers mats spectrum shows predominance of β phase, which offers the highest piezo-, pyro- and ferro-electric properties and it is particularly interesting for sensing applications.33 This result suggests that electrospinning process could have induced the formation of the β phase in the PVDF fibers mats.40,41
Fig. 2 FTIR spectra of (a) PPy particles, (b) electrospun PVDF, (b) PVDF/PPy_3, (c) PVDF/PPy_13 and (d) PVDF/PPy_23. |
The thermogravimetric curves of PPy particles and of electrospun mats of neat PVDF and PVDF/PPy blends are reported in Fig. 3. PPy particles present a main weight loss starting from 320 °C, which corresponds to the polymer chain degradation. On the other hand, PVDF shows weight loss at 473 °C which is attributed to the polymeric chain decompositions, with 5.50% of residue at 700 °C.42 The PVDF/PPy blends prepared with various amounts of PPy start to decompose at temperatures higher than that of pure PVDF and showed a single weight loss attributed to the thermal decomposition of both components. As a result, the maximum decomposition temperature shifts to higher values with increasing PPy concentration. This behavior can be attributed to the site-specific interaction between PPy and PVDF groups. Moreover, at 700 °C the residue obtained for the PVDF/PPy blends was higher than pure PVDF due to the presence of PPy and changes accordingly with PPy concentration in the blends.
Fig. 3 TGA curves of (a) PPy particles, and electrospun (b) PVDF, (c) PVDF/PPy_3, (d) PVDF/PPy_13 and (e) PVDF/PPy_23 mats. |
Fig. 4 shows the DSC thermogram of electrospun mats of PVDF and PVDF/PPy blends with various PPy weight fractions. An endothermic peak at 54 °C can be clearly detected on the thermogram of pure PVDF. This peak, which disappears in the immediate second heating run, has been already reported for PVDF with various possible origins: upper glass transition,43 reorganization within conformationally disordered α-crystals,44 molecular motions corresponding to an αc relaxation in the crystalline/amorphous interface,45 melting of paracrystalline domains.46 More recently, further insights on the nature of this annealing peak has been provided by Neidhöfer et al.47 An endothermic transition at 157.9 °C is assigned to the melting of α-form crystalline structure while the second endothermic peak at 167.9 °C can be related to the β-form of PVDF crystals. According to the literature,48,49 double melting transitions are frequently observed in PVDF due to its polymorphic structure. This behavior can be associated to the fact that the high electrical field and mechanical stretching typical of the electrospinning process can induce the formation of piezoelectric β-phases in electrospun PVDF fibers.50 For the electrospun PVDF/PPy blends obtained with different PPy content, the endothermic transition associated with α or αc phases is observed at around 60 °C and the melting temperature is located, at 158.8, 159.6 and 161.6 °C for blends containing 3, 13 and 23 wt% of PPy, respectively.
Fig. 4 DSC curves of electrospun (a) PVDF, (b) PVDF/PPy_3, (c) PVDF/PPy_13 and (d) PVDF/PPy_23 mats. |
Table 2 shows the melting temperature (Tm), the sample enthalpy of fusion (ΔHf) and the crystallinity (Xc) of electrospun PVDF and PVDF/PPy blends. It is worthwhile to note that neat PVDF displays a crystallinity degree higher than other values reported in the literature for PVDF electrospun mats.38,40 Additionally, the PVDF crystallinity is higher than electrospun PVDF/PPy blends, thus indicating that the addition of PPy particles can hinder the crystallization of PVDF chains.42
Sample | ΔHf (J g−1) | Xc (%) | Melting temperature (Tm) (°C) |
---|---|---|---|
PVDF | 82.9 | 79.0 | 157.9 and 167.9 |
PVDF/PPy_3 | 48.8 | 49.0 | 158.8 |
PVDF/PPy_13 | 61.7 | 59.0 | 159.6 |
PVDF/PPy_23 | 42.4 | 55.4 | 161.6 |
Storage modulus (E′) as a function of temperature for electrospun mats of neat PVDF and PVDF/PPy blends is reported in Fig. 5a. Over the entire temperature range, E′ values of neat PVDF are higher than those found for the blends over the whole temperature range. In addition, the storage modulus of the blends decreases abruptly when the PPy content increases. This effect can be attributed to the lower crystallinity of the PVDF/PPy blends and the presence of PPy in the electrospun mats which reduces the intermolecular dipole bonding of PVDF fibers, and hence, reducing the Young modulus.
Fig. 5 DMTA traces (a) storage modulus and (b) loss factor of electrospun mats of neat PVDF and PVDF/PPy blends. |
Loss tangent (tanδ) thermograms for electrospun mats of neat PVDF and PVDF/PPy blends are reported in Fig. 5b. The two transition temperatures at around −42 °C and 86 °C, corresponds to the glass transition and relaxation process associated with molecular motions of crystalline fraction, respectively, and are affected by the presence of PPy, with exception of the blend containing 23 wt% of PPy.51,52 The shift of Tg to higher temperatures can be tentatively explained by considering the decrease in the mobility of the PVDF chains due to the presence of PPy.53 Moreover, the intensity of the peak corresponding to the glass transition temperature reduces significantly as the amount of conducting polymer in the blend increases. However, the tanδ intensity and Tg of PVDF/PPy blends with 23 wt% are quite similar of those for pure PVDF mats. This behavior can be related to the presence of PPy particles or agglomerates, located almost exclusively outside the electrospun fibers.
Fig. 6a–c shows the variation of the electrical resistivity of PVDF and PVDF/PPy electrospun mats subjected to 25 consecutive compressive loading–unloading stress cycles. As expected, neat PVDF does not display a significant change of the electrical resistivity when the compressive stress is applied, since in piezoelectric materials it is not the resistance which is expected to change but the polarization charges and the induced voltage. On the other hand, the compressive stress has a remarkable effect on the electrical resistivity of the blends. When the system is subjected to a compressive stress (loading cycle), the resistivity value decreases several orders of magnitude as the pressure increases. Moreover, it is interesting to note that the initial resistivity value is almost completely recovered when the applied stress is removed. The minimum pressure value which the PVDF/PPy mats display detectable variation on the electrical resistivity is around 0.05 MPa, since the electrical resistivity reduces almost one order of magnitude. It is worthwhile to note that the PVDF/PPy blends with 3 and 13 wt% of PPy exhibit very wide electrical resistivity changes (ranging from 1016 down to 108 Ω cm and from 1017 down to 107 Ω cm, respectively), with reproducible response after 25 cycles. The sensitivity, i.e. variation on the electrical resistivity with the compression stresses, is much higher than the ones reported in the open scientific literature for conventional systems based on conducting polymers.1–5 However, when the conductive filler content is increased to 23 wt%, the response is markedly different. In fact, the electrical resistivity changes from 1014 up to 107 Ω cm in the first compressive ramp, due to the arrangements of fibers. However, after unloading and during the subsequent cycles, the electrical resistivity variation is limited to one order of magnitude. This behavior could be attributed to the high amount of PVDF/PPy agglomerates outside the electrospun fibers, as shown in FESEM images (Fig. 2d). After the first loading cycle the PVDF/PPy agglomerates touch each other, creating a preferential conductive pathway and consequently reducing the resistivity of the material. Since the agglomerates are not inside the fibers they cannot take advantage of the elastic recovery of the electrospun polymer network after load removal. Increasing the number of loading–unloading stress cycles the PVDF/PPy_23 sample presents a reduction in the variation magnitude of electrical resistivity due to the irreversible deformation.
As schematized in Fig. 7 the PVDF/PPy electrospun mats can be represented as a three-dimensional fiber network structure with randomly distributed fibers, with PPy particles located inside the fibers for blends containing up to 13 wt% of PPy. The microstructure was confirmed by TEM micrograph of fiber of PVDF/PPy_13, as showed in Fig. 7. TEM analysis indicated the presence of PPy particles embedded in the PVDF matrix. In this case, the average inter-fiber gap allows only relatively few contacts between the fibers and limited conducting pathways. Consequently, the electrical resistivity of the mat is high, with values quite similar to that of neat PVDF mat. Under compressive stress the inter-fiber gap is reduced, the contacts between fibers may form a conducting pathway. Moreover, the applied compressive stress can reduce the distance between the PPy particles inside the fibers, thus decreasing the electrical resistivity of the blends. Both phenomena are responsible for the large and reversible variations of the electrical resistivity observed of the electrospun blends under compressive stress.
Fig. 7 Illustrative scheme of the conducting pathway formation in electrospun conductive polymer blend without pressure and after applying an external pressure. TEM of PVDF/PPy_13 is also reported. |
Fig. 8a shows the dependence of the electrical conductivity on compressive stress during 25 consecutive loading–unloading cycles performed on the electrospun blend with a PPy content of 13 wt%. This blend composition was selected for a deeper analysis due to its higher sensitivity when compared with the others PVDF/PPy compositions investigated in this work. It is worthwhile to note that the loading–unloading process induce a very reproducible response which is not significantly affected by increasing the number of cycles. This behavior indicates a reversible organization of the fibers in the electrospun mat and, consequently, a restoration of the conducting pathway due to the absence of plastic deformation.
Fig. 8b shows the hysteresis effect on the stress–strain curves when the PVDF/PPy_13 electrospun mat was subject to 25 consecutive loading–unloading cycles. A residual irreversible strain can be observed on unloading after the first loading step probably related to the sample accommodation between the electrodes. However, from the 2nd to 25th cycle no variations of the hysteresis loops can be observed, thus excluding the presence of irreversible phenomena (such as plastic deformation) on the sample. The sensibility of the sensor can be quantified by its response (e.g. voltage, resistivity) over the stress applied. The sensibility of the PVDF/PPy_13 mats varies from 20 to 200 Ω cm per Pa, which is considerably higher than the one found for piezoelectric systems based on neat PVDF fibers (1.0–1.66 × 10−3 V cm−1 per Pa).27 However it is important to note that the pressure sensor based on a PVDF/PPy mats is designed to work under pressures greater than 0.02 MPa to 5 MPa, while piezoelectric sensor based on PVDF is sensitive in the range from 0.1 Pa to 12 Pa.
SEM micrographs were performed in the PVDF/PPy_13 after 5 and 25 cycles and reported in Fig. 9. Electron microscopy does not reveal differences in the morphology of the electrospun fibers before and after the compressive tests, which means that there are no plastic deformation phenomena within the 25 loading–unloading cycles. This observation support the mechanical results presented in Fig. 8b, confirming the reversibility of the sensor microstructure upon repeated loading–unloading cycles.
Fig. 9 SEM micrographs of electrospun mats of PVDF/PPy_13 after (a) 5 and (b) 25 consecutive compressive loading–unloading stress cycles. |
This journal is © The Royal Society of Chemistry 2014 |