1. INTRODUCTION
Contact electrification occurs when two dissimilar materials come into contact and
subsequently separate, resulting in charge transfer, primarily due to differences
in their electronic and surface properties[1-3]. For triboelectric charge generation, the conventional understanding is that a substantial
difference in the electron affinities and chemical compositions of the two materials
is required. However, recent studies have demonstrated that, contrary to previous
assumptions, contact electrification can also occur between identical materials and
is not exclusive to dissimilar materials[4-8]. This phenomenon has been observed for materials with variations in their surface
roughness and morphological structures. For example, smooth polymer surfaces tend
to acquire negative charges when in contact with chemically identical polymer surfaces
that are rough. This is primarily attributed to differences in the mechanical stress
experienced during the contact-separation cycles[6,
9]. Specifically, smoother surfaces undergo tensile stress during contact-separation
cycles. This tensile stress lowers the energy barrier for bond scission, resulting
in the formation of negatively charged fragments on surfaces. By contrast, a rougher
surface experiences compressive stress, thereby retaining its positive surface charge.
This recent finding of contact electrification between identical materials offers
significant potential for the development of energy-harvesting devices such as piezoelectric
nanogenerators (PENGs) and triboelectric nanogenerators (TENGs). This approach simplifies
material selection and fabrication processes by making it possible to control charge
polarity using structural alterations rather than compositional modifications. Among
the various materials investigated for energy-harvesting devices[10,
11], polyvinylidene fluoride (PVDF) has garnered significant attention because of its
unique combination of desirable properties, including exceptional mechanical flexibility,
chemical stability, and structural integrity[12-14]. PVDF is a promising triboelectric layer for TENGs. The presence of ordered permanent
molecular dipoles in its electroactive β-phase crystalline structure imparts high
electron affinity, which in turn significantly enhances charge transfer during contact-separation
motions[15,
16]. Concurrently, the pronounced piezoelectricity associated with this β-phase enables
efficient mechanical-to-electrical energy conversion in PENGs[17,
18]. Among efforts to increase the performance of PVDF-based PENGs, many studies have
focused on optimizing β-phase content and dipole alignment.
Despite extensive optimization of its β-phase content and dipole alignment, the performance
of PVDF-based PENGs remains constrained by intrinsic material limitations. The primary
constraint is the comparatively low piezoelectric coefficient (d33) of PVDF, which typically ranges from 20 to 30 pC/N[19,
20]. This value is approximately one order of magnitude lower than that of its ceramic
counterparts, such as lead zirconate titanate. Consequently, the electrical output
generated from ambient mechanical energy is often insufficient to power portable electronic
devices, which hinders their practical applications. Furthermore, conventional methods
employed to maximize the β-phase fraction, such as mechanical stretching, high-voltage
poling, and thermal annealing, may introduce microstructural defects[21,
22]. These imperfections can compromise the long-term mechanical integrity and operational
durability of materials. To address these shortcomings, alternative strategies that
incorporate nanofillers such as barium titanate[23], zinc oxide[24], and carbon nanotubes[25] into PVDF matrices have been explored. While these approaches have yielded incremental
improvements in output performance, they also introduce disadvantages, such as increased
fabrication complexity and diminished stability under cyclic mechanical loading due
to interfacial defects between the polymer matrix and nanofillers[26,
27]. Moreover, the addition of certain fillers can increase dielectric loss, creating
leakage current pathways that reduce overall energy conversion efficiency[28,
29].
In this paper, a novel hybrid nanogenerator (HNG) based on a monolithic PVDF nanofiber
(NF) architecture is proposed. This approach challenges the conventional understanding
that triboelectric effects are negligible between identical materials under mechanical
preloading conditions. By electrospinning and assembling alternating layers of large-
and small-diameter NFs, a volumetric laminate structure was fabricated that synergistically
integrated the intrinsic piezoelectricity with a triboelectric effect. In addition
to the piezoelectricity of PVDF, the enhanced performance of the device is driven
by two concurrent triboelectric mechanisms: (i) structural contact electrification
induced by the physical asymmetry between large and small diameter fibers, and (ii)
phase-induced contact electrification arising from differences in the electroactive
β-phase content between the two fibers, resulting in distinct electron affinities.
This volumetric energy generation mechanism is highly scalable, as the electrical
output increases proportionally with the number of layers owing to the increased triboelectric
effects, as confirmed by fast Fourier transform (FFT) analysis.
Furthermore, when operated in contact-separation mode, the laminate structure achieved
a higher output performance than the non-laminate PVDF NF-based TENG. By leveraging
these synergistic effects, significantly enhanced electrical output was demonstrated
in both the mechanical preload and contact-separation modes without the need for dissimilar
materials and fillers. More importantly, unlike previously reported PVDF-based HNGs
that rely primarily on externally driven contact–separation actuation[30-32], the proposed architecture enables simultaneous triboelectric and piezoelectric charge
generation even under strain-dominated actuation, resulting in hybrid energy harvesting
with improved durability and operational stability.
2. EXPERIMENTAL
2.1 Materials
PVDF powder (Mw ≈ 534,000 g/mol) was purchased from Sigma-Aldrich (St. Louis, MO,
USA). The solvents used for electrospinning, N,N-dimethylformamide (DMF) and acetone (ACE), were acquired from Duksan Reagent (Ansan,
Republic of Korea). Aluminum (Al) tape was supplied by Ducksung (Seoul, Republic of
Korea).
2.2 Solution preparation
PVDF solutions were prepared at two different concentrations: 10 and 18 wt%. The 10
wt% solution was prepared by adding 0.94 g of PVDF powder to 10 mL of a solvent mixture
consisting of ACE and DMF in a 6:4 volume ratio. Similarly, the 18 wt% solution was
prepared by adding 1.83 g of PVDF powder to 10 mL of an ACE and DMF mixture at a 7:3
volume ratio. Each solution was then stirred on a hotplate at 100°C and 300 rpm for
1 h to dissolve the PVDF powder.
2.3 Fabrication of the PVDF NF mat and HNG device
The prepared PVDF solutions were electrospun into NFs using an electrospinning/spray
system (ESR200; NanoNC, Seoul, Republic of Korea). To collect the NFs, an Al foil
was attached to a rotating drum collector operating at a speed of 10 rpm. The specific
electrospinning parameters were tailored for each solution concentration. For the
10 wt% solution, a 27-guage needle was used. The process was conducted with an applied
voltage of 15 kV, a solution flow rate of 20 µL/min, and a tip-to-collector distance
of 16 cm. For the 18 wt% solution, an 18-gauge needle was employed with an applied
voltage of 9 kV, flow rate of 40 µL/min, and tip-to-collector distance of 10 cm. After
electrospinning, the resulting NF mat was annealed in an oven at 80 °C for 2 h to
evaporate residual solvent. For device fabrication, the annealed NF mat was detached
from the Al foil and subsequently laminated onto Al tape, which served as the electrode.
Finally, a 1 × 1 cm2 square of Al tape was affixed to the tip of the mechanical pushing tester as a counter
electrode.
2.4 Characterization
The surface morphologies of the PVDF NFs were observed using field-emission scanning
electron microscopy (FE-SEM; JSM-6700F, JEOL, Tokyo, Japan). The diameters of the
NFs were measured from the FE-SEM images using ImageJ software (National Institutes
of Health, Bethesda, MD, USA). Fourier-transform infrared (FTIR) spectroscopy (Nicolet
iS50, Thermo Fisher Scientific, Waltham, MA, USA) was used to analyze the crystalline
structure and phases of the PVDF NFs. The frequency-dependent dielectric constants
of the NF mats were measured using an LCR meter (E4980AL; Keysight Technologies, CA,
USA). The output performance of the HNG was characterized by evaluating its response
under two different mechanical conditions using a pushing tester (JLPT-100; JunilTech,
Daegu, Republic of Korea).
First, to characterize the output in continuous contact mode, a cyclic deformation
force was applied. A mechanical preload of 2 N was set, and the force was modulated
up to 30 N at a frequency of 0.5 Hz.
Second, to measure the output in the contact-separation mode, the tester was configured
to maintain a separation distance of 5 cm by applying a cyclic force of 30 N at a
0.5 Hz frequency. For both measurement conditions, the signals were recorded using
a source measurement unit (SMU; 2612 B; Keithley Instruments, Cleveland, OH, USA).
For signal processing, FFT was performed using MATLAB 2025b. The analysis was conducted
at the fundamental frequency (0.5 Hz) corresponding to the mechanical testing frequency.
The resulting amplitude signals were normalized between 0 and 1.
3. RESULTS AND DISCUSSION
3.1 Fabrication and Structure of the HNG
The fabrication process and resulting structure of the HNG are schematically illustrated
in Fig. 1a. A volumetric multilayered laminate structure was fabricated using sequential electrospinning.
Specifically, PVDF solutions of two different concentrations (18 and 10 wt%) were
electrospun alternately to form a layer-by-layer assembly. The process began with
the electrospinning of the 18 wt% PVDF solution, creating a large-diameter NF mat,
followed by the deposition of a small-diameter NF mat from the 10 wt% PVDF solution.
This alternating deposition process was repeated to achieve the desired total number
of layers for the HNG.
Fig. 1. (a) Illustration of the fabrication process of the HNG, involving the alternating
electrospinning of 18 wt% and 10 wt% PVDF solutions to form a volumetric laminate
structure. (b) Operating mechanism, driven by the synergistic integration of piezoelectric
and triboelectric effects under an applied mechanical force.
The energy-harvesting capability of HNG originates from the synergistic integration
of the intrinsic piezoelectric and triboelectric effects, as depicted in Fig. 1b. When an external compressive force was applied, the layered PVDF NF mats mechanically
deformed, which activated the piezoelectric properties inherent to the PVDF. Concurrently,
triboelectric charges were generated at the engineered interfaces within the laminate
structure.
Specifically, two distinct types of interfaces were formed between the alternating
NF layers. When small-diameter NFs were deposited onto the underlying large-diameter
NF mat, they interpenetrated the network, creating a fixed interface with strong interlayer
adhesion. Conversely, when large-diameter NFs were electrospun onto a dense, small-diameter
NF mat, the significant size difference prevented interpenetration. This results in
weak interlayer adhesion and forms a slip interface that is prone to friction and
relative motion under mechanical stress, establishing it as the primary site for contact
electrification. These piezoelectric and triboelectric effects work together to produce
a hybrid electrical output. The detailed mechanisms behind the charge generation are
analyzed in the following sections.
The surface morphology of the electrospun PVDF NFs produced from the 10 and 18 wt%
solutions was characterized using SEM, as presented in Fig. 2a and 2b, respectively. A dense web-like network of NFs was obtained from the 10 wt% solution,
as shown in Fig. 2a, whereas significantly thicker fibers were produced from the 18 wt% solution, as
shown in Fig. 2b. This visual difference was quantified by analyzing the NF diameter distribution
(Fig. 2c, d). The NFs from the 10 wt% solution exhibited a narrow size distribution with an average
diameter of 0.15 µm, which increased substantially to 0.74 µm for the 18 wt% solution.
This increase was primarily attributed to higher solution viscosity and greater polymer
chain entanglement at higher concentrations.
These factors offer greater resistance to the electrostatic stretching of the polymer
jet, inhibiting its thinning before solidification and thus resulting in a larger
diameter[33-35]. The formation of thicker fibers was further controlled by optimizing other electrospinning
parameters, including lowering the gauge needle, shortening the tip-to-collector distance,
and reducing the applied voltage, as detailed in the Experimental Section.
The successful assembly of these distinct NF mats into a laminate structure was confirmed.
Fig. 2e presents a top-down view of the fixed interface, showing the uniform deposition of
small-diameter (10 wt%) NFs onto the underlying layer of the large-diameter (18 wt%)
NFs. The magnified view in Fig. 2f further reveals the interpenetrated network, which creates a high degree of mechanical
interlocking and strong interlayer adhesion.
Fig. 2. SEM images of NFs fabricated from (a) 10 wt% and (b) 18 wt% solutions. (c,d)
Histogram of corresponding diameter distribution of (c) 10 wt% and (d) 18 wt%. (e)
Top-down view and (f) magnified SEM images of the fixed interface, where the small-diameter
NFs are deposited onto the large-diameter NFs.
3.2 Crystalline Phase and Dielectric Properties
The crystalline phases of the fabricated PVDF NFs were investigated to determine their
electrical properties. The FTIR spectra of the NF mats produced at 10 and 18 wt% are
presented in Fig. 3a. Both spectra exhibit prominent absorption bands corresponding to the non-electroactive
α-phase (762 cm-1) and the electroactive β-phase (840 cm-1 and 1275 cm-1). The β-phase fraction (F(β)) was quantified based on the Beer–Lambert law using Equation (1)[36]:
where A
α and A
β represent the absorbances at 762 and 840 cm-1, respectively, and K
α (6.1×104 cm2/mol) and K
β (7.7×104 cm2/mol) are the corresponding absorption coefficients. The analysis revealed that the
18 wt% NFs possessed a higher F(β) (83.2%) than the 10 wt% NFs (73.4%).
An increase in PVDF concentration leads to higher viscosity and stronger macromolecular
chain entanglement, which modifies the degree of jet stretching during electrospinning.
More effective stretching in the higher concentration range promotes molecular chain
alignment along the jet axis, thereby enhancing the formation of the β crystal phase[37-40]. This stretching process forces the polymer chains into an all-trans (TTTT) conformation, thereby favoring crystallization into the polar β-phase over
the non-polar α-phase[35]. The higher β-phase fraction in the 18 wt% sample directly influences its dielectric
response. Because the β-phase contains aligned –CF₂– dipoles, a greater β-phase content
results in a higher density of permanent dipoles per unit volume. Under an alternating
electric field, these dipoles contribute to enhanced dipolar polarization, leading
to an increased dielectric constant[41,
42]. This interpretation was further supported by dielectric measurements performed on
solution-cast PVDF films. As shown in Fig. 3b, the film prepared from the 18 wt% solution exhibited a significantly higher dielectric
constant (6.91 at 10³ Hz) than that from the 10 wt% solution (4.73 at 10³ Hz). The
dielectric constant was evaluated in the frequency range of 1 kHz to 1 MHz as the
dielectric response is dominated by dipolar polarization within this frequency window[43].
Fig. 3. (a) FTIR spectra of the electrospun NF mats fabricated from 10 wt% and 18
wt% PVDF solutions. (b) Dielectric constant as a function of frequency for solution-cast
films prepared from the 10 wt% and 18 wt% PVDF solutions.
3.3 Output Performance of the HNG in Mechanical Preload
The electrical performance of the HNG device was evaluated under a mechanical preload
to investigate the piezoelectric and triboelectric effects. The output had three characteristic
configurations: single-component NF mats made from 10 wt% and 18 wt% solutions, and
12-layer HNG composed of alternating NF mats made from 10 wt% and 18 wt% solutions.
As shown in Fig. 4a, each device was subjected to a cyclic compressive force of 30 N with a constant
preload of 2 N. This preload condition ensures continuous contact between the counter
electrode and NF mat surface, thereby eliminating signals from contact electrification
at the electrode–NF interfaces. The output voltages and current densities are presented
in Fig. 4b and 4c, respectively.
The single-component mats with 10 wt% and 18 wt% NFs generated a cyclic electrical
output, attributed to the intrinsic piezoelectricity of PVDF. The 18 wt% NF mat exhibited
a slightly higher average peak-to-peak voltage (V
pp) of 0.037 V and average peak-to-peak current density (J
pp) of 0.31 nA/cm2 compared with the 10 wt% NF mat, which generated a V
pp of 0.028 V and J
pp of 0.19 nA/cm2. This enhancement is attributed to the higher β-phase content, as confirmed by the
FTIR analysis in Fig. 3a, which results in a stronger piezoelectric response. In contrast to the single-component
mats, the 12-layer HNG exhibited a dramatically enhanced output, reaching a V
pp of 0.22 V and J
pp of 1.54 nA/cm2.
Fig. 4. (a) Illustration of the experimental setup for the three distinct HNG materials
at the mechanical preload conditions. (b) Output voltage and (c) current density of
the HNGs depending on the materials.
This substantial increase demonstrated the successful operation of the hybrid energy-harvesting
mechanism. While the HNG generates piezoelectricity throughout all 12 layers, the
primary source of output enhancement is the introduction of triboelectricity at numerous
internal slip interfaces within the volumetric laminate structure. When mechanical
deformation occurs, these interfaces generate triboelectric charges via two concurrent
mechanisms. First, structural contact electrification arises from the topographical
differences between the large- and small-diameter NFs. When mechanical stress is applied,
the smoother surface of the large-diameter NFs experiences greater tensile strain
during contact, which lowers the energy barrier for bond scission and promotes the
formation of negatively charged polymer fragments[6,
7]. Consequently, a net negative charge accumulates on the large-diameter NFs, whereas
the small-diameter NFs become positively charged. Second, phase-induced contact electrification
provides an additional driving force for charge transfer. The large-diameter NFs possess
a higher β-phase content, resulting in a higher surface density of electronegative
CF2 groups. This enhances the effective electron affinity of the material, causing the
large-diameter fibers to accept electrons from the small-diameter fibers. Both mechanisms
cause the large-diameter fibers to become negatively charged upon contact. The resulting
triboelectric effects act in concert with the piezoelectric effects, leading to a
significantly amplified device output.
Notably, this enhancement is achieved under mechanical preload conditions, where the
actuation is predominantly strain driven. Despite operating in a mode typically used
for conventional PENGs, the proposed HNG exhibits substantially higher output performance,
demonstrating its superior efficiency for strain-based energy harvesting. This feature
makes the proposed HNG particularly advantageous in environments involving continuous
mechanical compression or vibration, where contact–separation motion is limited and
friction-induced wear must be minimized.
To demonstrate the scalability of the volumetric HNG structure, a series of devices
with increasing numbers of layers (2, 6, 12, 20, and 30) were fabricated and tested
under the same mechanical preload conditions. In addition, to further examine the
relative contributions of piezoelectric and triboelectric effects, the 30-layer device
was electrically poled under an electric field of 30 MV/m for 6 h, and its output
performance was compared with that of the unpoled device.
To minimize thickness-induced variations in device performance, the total electrospinning
time was fixed at 1 h for all devices. Instead, the deposition time for each individual
layer was adjusted according to the number of layers, while 10 wt% and 18 wt% solutions
were alternately electrospun to form the multilayer structure. In this study, the
multilayer structure was fabricated using a single-nozzle system with sequential solution
exchange. However, for scalable production, double- or multi-nozzle electrospinning
systems can be employed, allowing alternate operation of different solutions without
frequent replacement and thereby reducing processing time[44].
The output voltage and current density for various numbers of layers are presented
in Fig. 5a and 5b, respectively. The voltage and current densities monotonically scaled with the number
of layers. Increasing the layers from 2 to 30 resulted in a substantial enhancement,
with the V
pp increasing from 0.065 to 0.78 V and the J
pp increasing from 0.62 to 3.74 nA/cm2. As more layers are added, each layer introduces an additional slip interface, effectively
multiplying the number of sites for triboelectric charge generation. The summation
of the signals from these interfaces led to a substantial amplification of the output
performance.
Furthermore, comparison between the poled and unpoled 30-layer devices revealed no
significant change in the output magnitude, indicating that the piezoelectric contribution
is relatively minor under the present operating conditions. An FFT analysis was performed
on the voltage signals to determine the source of this performance enhancement (Fig. 5c). This approach is highly effective for distinguishing between piezoelectric and
triboelectric signals within a hybrid output. The spectral bandwidth is a key indicator
of triboelectric contributions because triboelectric signals introduce a wide array
of high-frequency harmonics owing to the nonlinear nature of contact electrification[45,
46]. To quantify this trend, each spectrum was normalized to its maximum peak intensity,
and the termination point (f3%) was defined as the frequency at which the signal intensity decreased to below 3%
of the maximum intensity. The spectral bandwidth was then calculated as the span between
this point (f3%) and the fundamental frequency (f0). The f3% value progressively shifted to higher frequencies as the number of layers increased,
reaching 7.5, 8, 10, 17.5, and 23 Hz for 2-, 6-, 12-, 20-, and 30-layer devices, respectively.
This corresponding increase in the spectral bandwidth is evidence of the progressively
dominant triboelectric contribution to the total hybrid signal as the number of layers
increases.
Fig. 5. (a) Output voltage and (b) current density as a function of the number of
layers. (c) Corresponding FFT spectra of the voltage signals for each layer configurations:
(i) 2 layers, (ii) 6 layers, (iii) 12 layers, (iv) 20 layers, and (v) 30 layers.
3.4 Output Performance of the HNG in Contact-Separation Mode
Fig. 6 presents an evaluation of the HNG performance in contact-separation mode, which represents
the typical operating configuration of a TENG. The working principle of the HNG in
contact-separation mode is illustrated in Fig. 6a. The cycle begins with the top electrode in contact with the HNG surface, at which
point the contact electrification establishes positive triboelectric charges on the
top electrode and an equal number of negative charges on the NF mat ((i) Contact).
As the top electrode starts to separate, a potential difference is induced between
the top and bottom electrodes owing to the distance between the positive triboelectric
charges on the electrode and the negative charges within the NF mat. These potential
differences drive electrons to flow from the bottom electrode to the top electrode
through the external circuit ((ii) Contact → Separation). When the maximum separation
distance is reached, charge transfer ceases as electrostatic equilibrium is established
((iii) Separation). As the top electrode approaches the HNG again, the potential difference
is reversed, forcing electrons to flow back from the top to the bottom electrode until
full contact is re-established, thus completing one alternating current cycle ((iv)
Separation → Contact).
The output voltage and current density for the single-layer (10 wt%), 12-layer, and
30-layer configurations are presented in Fig. 6b and 6c, respectively. Notably, the friction layer of all the configurations was a 10 wt%
NF mat to ensure a consistent triboelectric interface. Similar to the results under
preload conditions, an increasing trend in both the voltage and current density was
observed as the number of layers increased. Specifically, the output increased from
a V
pp of 61.4 V and J
pp of 0.25 µA/cm2 for the single-component 10 wt% NF mat to a V
pp of 161.49 V and J
pp of 1.3 µA/cm2 for the 30-layer HNG. This performance enhancement was attributed to the volumetric
design, in which the multiplication of the internal slip interfaces led to a greater
overall contact electrification effect.
The durability of the 30-layer HNG was tested over 10,000 contact-separation cycles
(Fig. 6d). The output voltage initially increased and then stabilized, which can be attributed
to the gradual accumulation of triboelectric charges within the porous NF network
during repeated contact–separation cycles. Under cyclic compression, charges generated
at the surface and internal slip interfaces redistribute into the inner layers of
the NF mat, resulting in progressive charge buildup until saturation is reached[47].
The power generation capability of each configuration was evaluated by measuring the
peak-to-peak load voltage (V
L) across a range of external load resistances (R
L). The corresponding power densities (P) were calculated using the equation P = V
L
2 / (R
L × A), where A is the contact area.
It was observed that as R
L increased, V
L increased for all three configurations (Fig. 6e). The maximum P was achieved at progressively higher R
L as the number of layers increased, occurring at optimal R
L values of 40, 60, and 70 MΩ for the 10 wt%, 12-layer, and 30-layer devices, respectively
(Fig. 6f). This trend indicated that the internal impedance of the HNG increased with the
number of layers. Achieving the maximum power transfer requires matching this internal
impedance with the external load, thus explaining the observed shift in the optimal
resistance.
Finally, practical energy-harvesting capability was demonstrated by charging capacitors
of varying capacitances (10, 47, 100, and 220 µF) through a full-wave bridge rectifier
(Fig. 6g). The time required to charge the capacitors to a given voltage increases with capacitance.
This series of experiments confirmed the practical energy-harvesting capability of
the proposed HNG in conventional contact–separation operation. Combined with its strong
output performance under mechanical preload conditions (Fig. 4), this dual-mode capability broadens its applicability across diverse mechanical
environments.
Fig. 6. (a) Working principle of the HNG in contact-separation mode. (b) Output voltage
and (c) current density of the 10 wt%, 12-layer, and 30-layer HNGs in contact-separation
mode. (d) Durability test of the 30-layer HNG over 10,000 cycles. (e) Peak-to-peak
voltage and (f) power density as a function of external load resistance for the three
distinct configurations. (g) Capacitor charging curves for the 30-layer HNG with various
capacitors.