1 Introduction

The growing demand for smart portable electronics and the batteries problem such as large size, the insufficient capacity, the danger of explosion, and the inconvenience of recharging, has led to the emersion of energy harvesters. Energy harvesters are able to provide long-term power supplies by converting available ambient energies from daily life usage to medical or military applications, such as pressure monitoring systems, implanted sensor nodes, unmanned vehicles, running military security systems, the importance of energy harvesting is increasing progressively. However, in the recent decade micro and nano energy harvesting has become more attention of the researchers [1,2,3]. Nanogenerators have several groupings and effects such as piezoelectric nanogenerator (PNG), triboelectric nanogenerator (TENG), thermoelectric generator (TEG), and pyroelectric generator (PYG) had been presented in the past decades for a variety of energy harvesting and self-powered uses. The problem with TENGs is that the electrical output is influenced by temperature and humidity [4]. pyroelectric materials are able to generate electrical energy with time-dependent temperature changes. All ferroelectric materials are pyroelectric, and all pyroelectric materials are piezoelectric, but the opposite is not correct. PVDF and its copolymers are used as pyroelectric energy harvesters due to their large pyroelectric constant [5]. Among all energy harvesting methods, electrical energy harvesting from mechanical energy sources is the most attractive and successful method since mechanical movements are independent of ambient conditions and also these ambient energies are mostly accessible [6]. Among the mechanical energy harvesters, piezoelectric materials are the most suitable and effective materials due to their versatility, connectivity, and integration, which are a good choice for using in smart textiles and flexible generators [7, 8]. The on-body demonstration for self-powered sensors is shown in Fig. 1. Common piezoelectric energy harvesters have an energy density of at least three times higher than those of other mechanical energy harvesters such as electrostatic and electromagnetic ones [9]. In order to use piezoelectric harvesters effectively, they should be fabricated from flexible materials instead of brittle piezoceramics.

Fig. 1
figure 1

On-body demonstration for self-powered sensors (This figure is reused from “Recent progress in human body energy harvesting for smart bioelectronics system” by Yang Zou, Lin Bo, Zhou Li, licensed under CC BY 4.0.) [10]

Full size image

Piezoelectric materials can generate an electric charge in response to applied mechanical stress; this phenomenon is called the direct piezoelectric effect, which is mostly used in energy harvesters. When a piezoelectric material is exposed under mechanical loading, there will be a replacement in the positive and negative charge centers in the material, which leads to an external electrical field [11, 12]. Piezoelectric effect is reversible, and the indirect piezoelectric effect occurs when an external electrical field causes a stretch or compress on the piezoelectric material [13,14,15]. Different types of piezoelectric materials are used in industry and technology, but the most popular and usable of them are PZT, BaTiO3, and PVDF [16,17,18,19]. Piezoelectric polymers are biocompatible, they have the mechanical flexibility and processability into different shapes and can directly be used as an active layer in energy harvesting applications [7, 20]. A commonly used piezoelectric polymer is polyvinylidene fluoride (PVDF) which is used in this study. PVDF is a suitable ferroelectric semi-crystalline polymer with four polymorphs α, β, γ, and δ (Fig. 2) [21, 22]. Although the α phase is the most common and stable phase, the β phase is of great interest due to the highest dipolar moment per unit cell and the largest piezoelectric properties in this phase [23, 24].

Fig. 2
figure 2

Chemical structure of different phases of polyvinylidene fluoride [25]

Full size image

Piezoelectric devices are the biocompatible and flexible devices with the ability of harvesting energy from both human in vitro and in vivo motion. The most recent research regarding different possible structures and designs of piezoelectric energy harvesters using for self- powered and wearable electronics are smart footwear, smart textile and smart skin. By contrast piezoelectric nanogenerators are known as a capable green energy harvesting technique [26]. In 2018, Talbourdet et al. investigated the differences between 2 and 3D woven fabrics from 100% optimized piezoelectric PVDF multifilament yarns. The textile structures were poled after the weaving process, and a maximum output voltage of 2.3 V was observed on 3D woven under compression by DMA tests. It was found that the 3D interlock design could improve energy harvesting due to the stresses of the multifilaments in the thickness [27].

In 2020, a thin polymeric film based on poly(vinylidene floride-co-hexafluoropropylene) (PVDF-HFP) and poly(methyl methactrylate) PMMA was used for energy harvesting by Polat. The results showed that blending with PMMA could increase the β-phase content and improved the heat stability. 84.7% increase found in the piezoelectric potential could be a favorable result for energy harvesting and sensors applications [3].

The list of developed piezoelectric generators in the similar works is reported in Table 1. It should be considered that different loading situations and sample dimensions lead to variety of output.

Table 1 The comparison of different polymeric piezoelectric energy harvesters
Full size table

The invention of piezoelectric generators provides a new procedure for powering portable electronic devices with low-energy consumption [36]. They can operate under stress or strain, such as arm twisting, knee bending, and expansion of the lungs and abdomen while breathing. Thus, a flexible piezoelectric substrate is a common choice for these environments [37]. It is clear that piezoelectric polymers (such as PVDF) are naturally flexible, and they are the preferred option for flexible harvesters. However, low power output is the main impediment of scavenging energy from a polymeric piezoelectric material that can be reached only in a specific frequency.

Furthermore, ambient energies usually exist at the low-frequency range and the low-frequency vibration sources such as human motions. It means that the electrical energy from these ambient sources has low amplitudes. Therefore, many approaches have been taken different strategies to improve the performance of piezoelectric energy harvesters at low-frequency [38, 39].

Piezoelectric materials are mostly bonded on a substrate and built in a rectangular cantilever shape. The more piezoelectric material cannot guarantee the efficiency of a harvester, and many studies have been done to improve the output power by modifying the geometry and the structure of energy harvesters [40,41,42]. The number of the piezoelectric layers and the type of electrode connection are other important parameters that influence the performance of an energy harvester [43]. It is reported that a multilayer bimorph energy harvester produces more power output at the same frequency and total thickness [44,45,46]. Other researches conducted in this field showed that, by increasing the number of layers, a small reduction of power output occurred, but the thickness of electrodes can explain this reduction. The theoretical analysis omitted the increases of electrode thickness, leaving the material volume unaffected [47]. The main factor of designing the structure of a piezoelectric generator is to maximize the applied stress or strain of an external load on the piezoelectric layer [48, 49]. This point is specifically crucial for a flexible piezoelectric generator. Because of the inherent softness of the matter, especially while bending, the stress or strain response may be concentrated which can, in turn, limit the amount of the electricity produced by the piezoelectric matter [50, 51].

There have been different endeavors to achieve high power output, such as different designs, nonlinear methods, optimization techniques, and harvesting materials. In some works, the effect of electrodes on a membrane energy harvester was considered. The results showed that electrode coverage of 60% revealed the best electromechanical coupling [29].

Optimizing the thickness was launched by Paquin and Amant, who suggested a semi-analytical model based on the Rayleigh–Ritz approximations [47]. In 2014, Zhao and You designed a 1 mW shoe-embedded PVDF energy harvester. For higher power output, the authors accumulated eight layers of PVDF films together and sandwiched them between two wavy surfaces. Under quick walking at ~ 1 Hz, a peak-to-peak voltage of 136 V, a peak power of 4 mW, and average power of 1 mW were obtained [9].

In 2015, a pair of PVDF-PI composites was assembled back to back to form a curved PVDF energy harvester by Jung et al. An arc-shaped polyimide (PI) substrate with a thickness of 0.2 mm was used and on each side of it, two 0.1 mm thick PVDF films were attached. In this formation, while loading, PVDF films experience bending stress without stretching. The sample was fabricated with four 70 × 40 mm2 PVDF films, which generated an averaged open-circuit voltage of 25 V at 0.5 Hz. The high efficiency of this harvester was due to the large surface area (multiple layers) and the high bending stress of the curved structure [5]. Furthermore, there were several examples of using non-piezo materials in energy harvesters. A "3D spacer" was demonstrated with the silver-coated polyamide multifilament yarn layers electrodes, and the spacer yarns were high β-phase PVDF monofilaments which provided a maximum power density of 5.07 μW/cm2.

Therefore this study attempts to present a multilayer hybrid structure including both piezoelectric (electrospun PVDF micro-nano fibers) and non-piezoelectric layers, called passive layer in this article, that has the required flexibility and at the same time, it produces proper output. The microstructural, crystalline structural, mechanical and thermal properties of the electrospun PVDF were investigated using the scanning electron microscope, X-ray diffraction, Fourier transform infrared spectroscopy, Instron instrument, and differential scanning calorimetry, respectively. An instrument called PiezoTester was manufactured for piezoelectric properties evaluation and electrical output measurement. The influence of the passive layer factors such as thickness, shape, and material on the electrical output was studied under bending and tapping load. The bending modulus of the prepared harvesters was measured to find a relation between bending modulus and electrical output. The difference between this work and the other reports is that they were based on improving the piezoelectric materials and output. However, this study aims to develop a new structure to improve the performance of the harvesters. In order to increase the power output, the mechanical damping effect of energy harvesters should be minimized.

2 Experimental section

2.1 Materials

Polyvinylidene fluoride polymer [PVDF; molecular weight (mW) 230000] was purchased in granule form from Halopolymer Company. Acetone (99%) and N,N-dimethylformamide (DMF, 99%) were purchased from Merck Company. Aluminum foil was used as electrodes. Cellulose, polyester, and aluminum layers were used as passive layers with different thicknesses.

2.2 Electrospinning

To produce piezoelectric fibers, electrospinning equipment with a rotary drum was used (Fig. 3). PVDF solution (26 wt/wt%) was prepared by dissolving PVDF granule in an acetone/DMF solvent mixture (2/3 wt/wt) and then the mixture was stirred with a magnetic stirrer for 2 h at 60° C. Finally, a clear solution was obtained.

Fig. 3
figure 3

Schematic view of the electrospinning setup used to fabricate PVDF nanofiber layers

Full size image

Nanofibrous layers were fabricated using the FNM© electrospinning system. Feeding rate, applied voltage, and the collector rotating speed were selected according to the previous studies [7, 52]. Based on the aforementioned proper conditions, the spinning distance (collecting distance) was considered 18 cm. A syringe with a volume of 1 ml and a needle with a gage of 22 (needle length 34 mm, inner diameter 0.4 mm and outer diameter 0.7 mm) were used in electrospinning. The spinning solution was loaded into a syringe and pumped with the flow rate at 0.5 ml h−1. An 18 kV voltage was applied between the needle and the drum collector.

2.3 Preparation of piezoelectric devices

After fabricating the fiber layers, the samples were prepared in combination with the passive layer, according to Table 2 and Fig. 4, for bending and electrical output tests. The samples for electrical output tests were prepared in two different forms.

Table 2 Sample naming based on thickness and used material
Full size table
Fig. 4
figure 4

Schematic view of the structure of the samples. The thickness of a nanofiber layer and embedded aluminum foil is 50 micron

Full size image

The specimens were prepared in 2 × 2 cm for tapping tests and in 1 × 5 cm for bending tests. Cardboards in different thicknesses were used as available cellulose layers, PET films were used as polyester layers, and aluminum foils were used as aluminum layers. The reason for using the aluminum layer as one type of passive layers was to investigate the influence of a conductive passive layer on the performance of the harvester.

2.4 Characterization

Shirley fabric thickness tester (mode: SDLO 34 with 0.01 mm accuracy) was used for measuring the thickness of produced passive layers based on ASTM D1777-96. To investigate the morphology of the produced micro-nano fiber layers and to ensure their uniformity, a scanning electron microscope (SEM, Philips Model XL30) was employed. The diameter of micro-nano fibers was measured by Image-J software. One hundred fibers were measured to calculate the mean fiber diameter for each sample. For crystal structure analysis, X-ray diffraction (XRD) was used (Equinox3000, INEL Company—Cu, Kα: 0.154 nm). Differential scanning calorimetry (DSC) was used to study the thermal properties and crystal structure of the electrospun micro-nano fibers using a DSC2010 TA instrument Co. Fourier transform infrared spectroscopy (FTIR) analysis was carried out on the layers by FTIR spectrometer (NEXUS 670, Nicolet Company). The bending behavior of the samples was measured by an Instron 5566 Universal Testing System instrument according to ASTM E855 standard, at the displacement rate speed of 1 mm min−1. For this purpose, suitable and movable support was first designed and manufactured according to the above standard and then mounted on the machine (Fig. 5).

Fig. 5
figure 5

A view of manufactured support used to measure the bending rigidity of the samples

Full size image

A homemade instrument called PiezoTester system was used (http://ffm.aut.ac.ir) to evaluate the piezoelectric properties and to measure the electrical output of the harvester [53]. This evaluation was performed in both tapping and bending modes. To do this test, from each sample, 5 specimens were evaluated and the mean value was reported.

3 Result and discussion

Since the same electrospun layer was used for all samples, the active layer parameters were discussed first.

3.1 The electrospun fiber layer morphology

Achieving a uniform piezoelectric property comes from the production of uniform and beadless micro-nano fibers. Therefore, uniform production of micro-nano fibers must be ensured. As can be seen in the SEM image shown in Fig. 6, the fabricated micro-nano fiber layers were homogeneous and beadless, so it can be stated that the electrospun fiber layer has the required uniformity to be used in a piezoelectric harvester.

Fig. 6
figure 6

The SEM image of a prepared sample. Fabrication parameters are: Distance: 18 cm, Applied voltage: 18 kV, Feed rate: 0.5 ml/h and Rotating speed: 400 rpm. The average diameter of the nanofiber is 620 nm

Full size image

3.2 Crystallinity

Among all crystalline phases in PVDF polymer, the beta phase is the most appropriate one for piezoelectric properties. The electrospun fiber layer was first subjected to DSC thermal analysis to determine its crystallinity, and then based on the acceptable range in the related articles; further evaluations were carried out. To calculate the degree of crystallinity for each sample, Eq. (1) was used. XC indicates the degree of crystallinity of the sample and ΔHLit is the melting enthalpy for the state where the material is 100% crystalline in the beta phase [54].

$${X}_{\mathrm{c}}= \frac{\Delta {H}_{\mathrm{m} (\mathrm{Sample})}\times 100}{{\Delta H}_{\mathrm{lit}}} , (\Delta {H}_{\mathrm{lit}}=103.4 \frac{j}{g})$$
(1)

Figure 7 shows the DSC spectrum for the electrospun mats. By measuring ΔH (with the help of TA Universal Analysis software) and placing it in the Eq. (1), the percentage of crystallinity of the fiber layer was calculated to be about 51.4%, which is in the appropriate range.

Fig. 7
figure 7

The DSC test diagram of the fabricated layer. ΔHm is 53.26 J.g−1

Full size image

After thermal analysis, FTIR spectroscopy was used to investigate the structure of the produced micro-nano fibers and calculate the beta phase crystallinity ratio of the entire crystal structure. The diagram of the performed test on the electrospun fibrous mat is shown in Fig. 8. To determine the crystallinity of the beta phase in each sample, the absorption spectral peaks of the alpha and beta phases were evaluated at 766 and 840 cm−1, respectively. The beta phase of the crystal was calculated from Eq. (2):

Fig. 8
figure 8

FTIR spectroscopy of the PVDF micro-nano fiber layer

Full size image
$$f\left(\beta \right)=\frac{{X}_{\beta }}{{X}_{\beta }+{X}_{\alpha }}\times 100= \frac{{A}_{\beta }}{1.26{A}_{\alpha }+{A}_{\beta }}\times 100$$
(2)

where Aα and Aβ are absorptions at 766 and 840 cm−1 for alpha and beta phases, respectively, and Xα and Xβ are crystallinity degrees for alpha and beta phases, respectively [55]. These values were extracted using OMNIC software and placed in Eq. (2). As shown in Table 3, 87% of the crystalline part is in the beta phase.

Table 3 FTIR spectroscopy results
Full size table

X-ray diffraction (XRD) was used as another crystal structure test to confirm and monitor the results obtained from FTIR and DSC experiments. As mentioned, the piezoelectric property of polyvinylidene fluoride polymer, results from the beta phase of its crystal structure. The presence of a spectral peak at an angle of 2θ = 20.5 in the X-ray diffraction spectrum of this polymer indicates the beta phase in its crystal structure. Also, the presence of a spectral peak at an angle of 2θ  = 18.5 indicates the alpha phase of the crystal structure [54]. As shown in Fig. 9, the spectral peak intensity at 20.5° is much higher than the spectral peak intensity at 18.5°, indicating that the beta phase is formed in the electrospun fiber layer. This is in confirmation of the results of the DSC and FTIR experiments.

Fig. 9
figure 9

XRD diagram of the micro-nano fibrous layer

Full size image

The crystallite size was estimated from the Scherrer Eq. (3) [56],

$$D=\frac{K\lambda }{\beta \mathrm{cos}\theta }$$
(3)

where D is the average crystallite size, λ is the x-ray wavelength, β is the width of the x-ray peak on the 2θ axis, normally measured as full width at half maximum (FWHM) after the error due to instrumental broadening has been properly corrected (subtraction of variances), θ is the Bragg angle, and K is the so-called Scherrer constant [57]. The constant K has been determined to vary between 0.89 and 1.39, but it is usually taken as close to unity, the assumption that K = 1.0 is generally justifiable [56]. At the 2θ value of 20.5, calculation has produces crystallite size is 5 nm.

All the above experiments proved that the fabricated layers have proper piezoelectric properties to consider them as active layers in the energy harvester devices. In the next step, the effect of passive layers on bending rigidity and electrical output was evaluated.

3.3 Bending rigidity

In order to study the effect of the passive layer on bending strength and flexibility of the produced specimen, a three-point bending test was performed based on the ASTM E855 standard. In this experiment, the bending modulus and the stress–strain response of the material were determined under bending loading. During the experiment, there was a compressive stress in the upper surface (above the neutral axis) and tensile stress in the lower one. Figure 10 shows a diagram of the sample in which the vertical axis represents the force and the horizontal axis represents the displacement.

Fig. 10
figure 10

Bending load versus displacement diagram of samples

Full size image

The bending modulus of each sample was calculated from diagrams (the slope of the first 5% of the diagram) and shown in Table 4 and Fig. 11. As can be seen, by adding a passive layer, the bending modulus of the samples was increased. This improvement in modulus increased with raising the passive layer thickness so that the modulus of S4 is more than 6 times higher than that of S1. In the samples with the same thickness of the passive layer, the aluminum layers had the highest modulus increase and the samples with the cellulose layer had the lowest modulus growth.

Table 4 Bending modulus of samples
Full size table
Fig. 11
figure 11

Variation of bending modulus in different samples

Full size image

In other words, adding a passive layer can reduce flexibility. This decrease in flexibility of the nanogenerator device is in its maximum for the S4 and then S8. ANOVA test was performed and the results showed that this difference between the flexibility of samples is significant (Supplementary data—SD1).

3.4 Electrical output

Due to the direct piezoelectric effect, if mechanical stress is applied to a piezoelectric material, an electrical response is obtained. For this reason, one of the important methods to evaluate the piezoelectric properties and performance of an energy harvester is to apply mechanical stress to the sample and measure its electrical output [55]. The experiment was performed on the samples in two different ways. In the first method, the specimens were subjected to a tapping load with a given force and frequency, and their electrical output was measured. In the second method, the specimens were subjected to a periodic bending force with a 5 mm displacement and their output was recorded. Since the ultimate goal is to use these harvesters in smart textiles and especially in garments that the applied load is in bending mode, the second method was considered.

3.4.1 Tapping mode of evaluation

To evaluate the piezoelectric properties of the manufactured samples, their electrical output was measured under a specified load and at a specified frequency. For this purpose, 5 specimens from each sample were prepared in 2 × 2 cm, with the thickness and arrangement mentioned in Table 2, and they were subjected to a 2.65 N tapping load with a frequency of 5 Hz. The average output value of each sample is shown in Table 5 and Fig. 12. ANOVA test was performed and the results showed that the data were significant (Supplementary data—SD2).

Table 5 Piezoelectric response of samples under tapping mode loading
Full size table
Fig. 12
figure 12

Electrical output diagram under tapping mode loading

Full size image
3.4.1.1 Passive layer material

As shown in Fig. 12, adding a passive layer causes a noticeable change in the electrical output of the fabricated nanogenerator device. Nevertheless, this change varies depending on the material of the passive layer. By comparing S1 to S4, it can be claimed that the amount of power output increased with the addition of the passive layer. However, in S5 compares to S7 and S6 compares to S8, adding the passive layer has a different trend so that in S5 and S7, the output is even lower than S1. As the thickness increases from 120 to 420 microns, the output value is increased in all three passive layers. The reason for the decrease in the electrical output in S5 and S7 can be attributed to the passive layer mechanical energy absorption, which is caused by the applied force. So less energy is transferred to the lower piezoelectric layers of the active layer and then the electrical output is reduced. In other words, it can be claimed that in practice, only one layer of nanofibers is effectively active in generating electrical energy. To investigate this claim, S1-1 with the same structure as S1 but with a monolayer of nanofiber was produced and tested. As can be seen in Fig. 13, the outputs of S5 and S7 are up to 20% higher than that of S1-1 and are not significantly different while their active layer content is 2 times higher. This result can be considered as a confirmation of the above claim that only one layer of nanofibers is effectively active in generating electrical energy. This effect is not observed in the samples with the aluminum passive layer because not only they are elastic, but also they have very low-energy absorption and are able to transfer energy from one side to the other. ANOVA test was also performed and the results showed that the data were not significant, confirming the damping claim (Supplementary data—SD3).

Fig. 13
figure 13

The effect of the passive layer material on electrical output

Full size image
3.4.1.2 Passive layer thickness

As can be seen in Fig. 12, in all 3 passive layer types, as the thickness of the passive layer increases, the electrical output increases and the sample outputs follow the same trend for all 3 types. In other words, it can be stated that by increasing the thickness (in specific material), the electrical output increases.

An essential point in the trend of output changes can be seen in the comparison of S6 and S8. As explained in the previous section, by adding a passive layer, the applied force is absorbed by the passive layer and is not transferred entirely to the bottom layer. However, with increasing the layer thickness, which is associated with increased energy absorption (increased viscoelastic effect), the output value is increased. At first sight, this is inconsistent with the above claim. This issue can be justified by two different perspectives, which are based on the capacitive property and device measuring mechanism. These two explanations are discussed as follow:

3.4.1.3 Explanation based on the capacitive property

As shown in Fig. 4, the samples are in fact capacitors consisting of two conductive aluminum plates and a multilayer dielectric. The following equations generally describe capacitors:

$$C = q/V$$
(4)
$$C = \varepsilon A/d$$
(5)

where C is the capacitance, q is the charge stored in the capacitor, V is the voltage across the capacitor, A is the common plane of the two capacitors, d is the distance between plates, and ε is the dielectric coefficient of the non-conductive layer between the two plates. Since the thickness of the piezoelectric nanofiber layer is the same in all the samples and assuming the applied force as constant, the induced charge in the samples can be considered the same (q1 = q2).

On the other hand, the dielectric coefficient is the same in the samples with the same passive layer (ε1 = ε2). All samples were also prepared in a similar size (A1 = A2). With these points in mind, it can be written:

$$\stackrel{(4)}{\Rightarrow }\frac{{C}_{1}}{{C}_{2}}=\frac{{V}_{2}}{{V}_{1}}$$
(6)
$$\stackrel{(5)}{\Rightarrow }\frac{{C}_{1}}{{C}_{2}}=\frac{{d}_{2}}{{d}_{1}}$$
(7)
$$\stackrel{(6),(7)}{\Rightarrow }\frac{{V}_{2}}{{V}_{1}}=\frac{{d}_{2}}{{d}_{1}}\Rightarrow {V}_{2}=\frac{{d}_{2}}{{d}_{1}}{V}_{1}$$
(8)

According to Eq. (8), with increasing thickness, the potential difference between the two samples increases. In fact, assuming constant load created in the samples by increasing the passive layer thickness, the potential difference of the samples increases.

3.4.1.4 Explanation based on the device measuring mechanism

In the PiezoTester method, as shown in Fig. 14, the force sensor is located precisely below the sample and measures the force applied to the sample. On the other hand, as mentioned above, the applied force is the same for all the samples and equals 2.65 N. It is conceivable that in this measurement system, the amount of force is not predetermined and the applied value is measured after applying force to the sample by the below sensor. When it is stated that the force is 2.65 N, it means that the force applied to the sensor is 2.65 N.

Fig. 14
figure 14

Schematic view of PiezoTester

Full size image

Since the samples have viscoelastic properties, the actual applied force is higher than this. The more viscoelastic the sample, the greater the difference between the measured force and the actual applied force is. When the passive layer thickness increased or, in other words, the viscoelastic effect of the samples becomes more significant, it is necessary to give more force to the samples with thicker passive layers than to the thinner layer samples so that the sensor can sense a constant force of 2.65 N. The higher the force, the more mechanical energy enters the upper piezoelectric layer, which results in the higher piezoelectric output. In fact, the increase in the output observed in S8 and S6 compared to S7 and S5, and even S1 is due to the increase in the applied force to the upper piezoelectric nanofiber layers during the measurement.

3.4.1.5 Effect of passive layer geometry on the electrical output

In order to investigate the effect of the geometry of the passive layer on the electrical output of the piezoelectric devices, S2-2 was produced using a netted aluminum foil to place between the two piezoelectric layers similar to the S2 layout, and its electrical output was evaluated (Fig. 15).

Fig. 15
figure 15

The netted aluminum layer used as a passive layer

Full size image

As can be seen in Fig. 16, the spectral peak intensities of each sample are approximately the same and the outputs are in equal magnitude. Therefore, it can be stated that passive layer with the selected geometry has little effect on the electrical output of multilayered piezoelectric devices. ANOVA test was performed and the results showed that the data were not significant (Supplementary data—SD4).

Fig. 16
figure 16

Electrical output spectra of sample S2 and S2-2

Full size image

3.4.2 Bending mode of evaluation

Since one of the most important applications of energy harvesters is the use in smart textiles and bending is one of the major deformations available in this application, piezoelectric property evaluation of the samples was also performed in bending mode. For this purpose, 5 samples, 5 × 1 cm, were produced with the arrangement mentioned in Table 1. The average electrical output of the samples is shown in Table 6.

Table 6 The average electrical output of the samples under bending loading
Full size table

Since each bending period consists of two loading and unloading stages, electrical output was observed at each step (Fig. 17).

Fig. 17
figure 17

Electrical output spectra under bending loading

Full size image

Figure 19 shows the mean electrical output under bending. As can be seen, by adding a passive layer, the electrical output increases in all devices with three types of passive materials. As the thickness of the passive layer increases, the electrical output increases. The reason for this can be attributed to the increase in the bending modulus of the samples. Another factor that increases the output is the increasing distance of the piezoelectric layer from the neutral axis with increasing the thickness. Every material under bending is divided into two parts (Fig. 18). In the upper part from the neutral axis, the structure is under a compressive force, and in the lower part from the neutral axis, the structure is under a tensile force. The higher the distance from the neutral axis, the greater effect of forces is expected. As the thickness of the passive layer increases, the distance from the neutral axis increases, and this factor extends the piezoelectric output.

Fig. 18
figure 18

The behavior of the structure under bending loading

Full size image

Besides, checking the Fig. 19 shows that the electrical output under loading is higher than the electrical output under unloading. The reason for this can be attributed to the fact that loading takes place under force, and unloading takes place inactively and without any external force. In this case, the ability to return to the original state is important.

Fig. 19
figure 19

The electrical output of samples under bending load

Full size image

As shown in the diagram, in the sample consist of polyester layer, due to better reversibility, the unloading output is higher than the ones made from cellulose or aluminum. However, in the aluminum sample, since the sample experiences plastic deformation after loading (no returns), the unloading output is smaller. ANOVA test was performed and the results showed that these results were significant (Supplementary data—SD2).

4 Conclusion

Nowadays, portable energy harvesters with extensive life span have grown for a wide range of applications. It is very likely that batteries will be removed from miniature devices in near future. This study introduced a new structure which can be directly integrated in potential wearables to charge portable electronics through body movements or even be used as a flexible and reliable power supplies. It is clear that the output voltage of the PVDF piezoelectric devices depends on the volume of material, and the amount of applied stress, the two factors that should be considered when designing a flexible piezoelectric assembly. On the other hand, the key aspect of designing a piezoelectric harvesting structure is to maximize the stress or strain created on the piezoelectric substrate by applying an external force. When the external force is applied in the middle of a PVDF film without a passive layer, it pushes the surface straight down, causing it to locally bend due to its large flexibility. However, the hybrid structure enables the PVDF to uniformly increase the applied stress on the entire surface of the material. In this status, the amount of electricity produced by the material is limited.

The results showed that under the tapping state, the electrical output of PVDF nanogenerator devices is highly dependent on the materials of the passive layer. If the passive layer is a high-energy-absorbing viscoelastic material, the addition of this layer has adverse effects, which decrease the electrical output comparing with the low-energy-absorbing materials.

Nevertheless, if a rigid layer capable of transmitting force is used, the addition of the passive layer increases the electrical output, and increasing its thickness leads to higher output. However, the bending test results showed that the use of such a layer has a limitation on device flexibility. The result of electrical output under bending was slightly different. These results showed that the addition of a passive layer increased the electrical output in all three materials and all thicknesses. As the thickness of the passive layer rises, the output also increases. Besides, if the sample is produced from a flexible and elastic passive layer, its output is similar under loading and unloading, and the output has a higher uniformity. Finally, it can be stated that the passive layer is more suitable in the harvesters under bending. It was concluded that adding aluminum, cellulose and polyester as passive layers could increase the electrical output about 4.7, 3 and 4.2 times more than the sample without any passive layer, respectively. In this case, the use of elastic passive layer is preferable comparing with the cellulose and aluminum layers. In the case that the nanogenerator is subjected to a tapping loading, the use of the passive layer is not justified and is only recommended if a rigid layer is used.

Thus, it seems that the number of applications will be increased in future. Efforts should be done to improve the performance of energy harvesters by fabricating new piezoelectric materials and composites. Optimizing the conditions to attain a high power output and a wide working bandwidth are the two goals which should be considered more in the future.