PAN/MXene: a potent piezo-gen and cogent piezoelectric separator for self-chargeable supercapacitors

Abstract

All-in-one self-charging devices with integrated storage systems have become one of the bottlenecks of research at present. Amidst a worthwhile contribution to this direction of research, we present a self-charging piezo-supercapacitor device (SCPS) consisting of carbonized polyacrylonitrile/MXene (C-PMX) nanofiber electrodes, a PMX nanofiber separator, and PVA/KOH as a gel electrolyte. First, the energy harvesting properties of PMX, followed by self-charging energy storage, were studied discretely. The piezoelectric nanogenerator (PNG) with PMX-3% showed the highest output voltage, short-circuit current, and power density of 9.9 V, 1124 nA, and 9.6 μW cm⁻², respectively. Consequently, the SCPS offered a device-specific capacitance of 100.5 F g⁻¹, along with an energy density of 5 Wh kg⁻¹ at a power density of 300 W kg⁻¹ and 92% capacitance retention. Further evaluation of self-charging was carried out with compression and tapping. Interestingly, the SCPS-CF can be charged up to 250 μV in just 0.2 s, while the discharge time takes about 1.5 s. Moreover, the influence of piezo-separators in piezo-electrochemical conversion and rectifications with SCPS-NF is discussed in detail. Thus, a new insight into the integration of harvesting and storage for future self-powered electronics has been brought in.

---

New concepts

The synchronization of distinct energy conversion and storage mechanisms relating to the piezo-electrochemical effect requires in-depth exploration. In addition, the utilization of nanofiber materials as different components of an integrated device has not yet been explored. Also, from the electrolyte perspective, acid-based electrolytes have mostly been studied for self-charging systems; hence, we experimented with a basic electrolyte for integrated performance. In this study, we utilized PAN/MXene nanofibers as a dielectric layer for piezo-devices, piezo-separators, and electrodes in supercapacitors. Interestingly, the carbonization process resulted in good conductivity in the PAN/MXene nanofiber electrodes. In addition, the high surface area and electrical and electrochemical properties of MXene further improved the device performance. Interestingly, we explored the potential of PAN/MXene as storage electrodes with carbon felt and nickel foam for real-time application along with providing rectification and flexibility. Thus, this research presents a promising and facile method for the utilization of piezoelectric nanofiber materials performing harvesting as well as storage functions in the same device.

1. Introduction

Lately, the incremental demand for the generation of electricity has become more vital for powering electronics.^1–3 Unfortunately, most of the existing technologies for the generation of electrical energy are frequently fuel-consuming.⁴ Hence, an already well-known technique that converts low-frequency vibrations into useful energy, the so-called “piezoelectric energy,” is of great interest.⁵ Among various energy sources, the aforementioned is more pervasive and convenient for the conversion of mechanical energy into electric power.^6–10 This mechanism is wholly entrenched in the intrinsic polarization of a material rather than an external driving source requirement. Moreover, it is highly durable and reliable and exhibits higher power and voltage output along with sensitivity to minute strains. Additionally, the integration of this harvested energy within a storage system paves the way for self-chargeable devices.^11–14 On account of the energy predicament, researchers focus on bringing a new approach to coalesce energy harvesters and storage through extrinsic or intrinsic modes to attain self-powered systems that can be creditable for different applications from microscale to macroscale.^15,16 Self-charging energy storage systems use applied mechanical energy to store electrochemical energy. As such, intrinsic integration is advantageous compared to extrinsic integration, as the latter requires a complex power management system with bulky circuitry, diodes, and bridge rectifiers for impedance matching to connect both systems.¹⁷ Moreover, it incurs additional manufacturing costs. Hence, for the sake of reducing production cost, the weight of the hybridized device, and power loss intrinsic to integration, it is a ready-to-go option. Hitherto, self-chargeable devices that can harvest, convert, and store as intrinsic are highly engrossing for research and product development due to their multi-functionality.^18,19 Successively, piezo-separators (PS) pave the way for next-generation fast and self-chargeable devices by making use of their in vivo energies. The device design of SCPS comprises two supercapacitor electrodes and a piezo-polymer separator covered with gel electrolyte (aqueous or ionic). Further, the electrodes of supercapacitors are advantageous over battery electrodes due to the high power density and fast charging and discharging as in cooperation with PS minute strains.

To date, there have been many reports on enhancing the energy and power density of supercapacitors. Also, the self-discharge in supercapacitors via charge distribution due to desorption of ions, leading to decay of potential, faradaic reaction, and ohmic leakage, is a major lacuna in these systems.²⁰ On the other hand, self-charging efficiency or other standards required for real-time applications and the contribution of electrodes and electrolytes in correlation with the performance of SCPS are not explored. Specifically, the synchronization of distinct energy conversion and storage mechanisms relating to the piezo-electrochemical effect is not completely understood. Hence, the incorporation of a PS in the perspective of improving the SCPS performance and the utilization of the same material as an electrode with enhanced conductivity are explored. The piezoelectric and electrochemical properties of piezoelectric separators, then, are both critical metrics in SCPS.

In spite of the classification of piezoelectric polymers into amorphous and semi-crystalline, the mechanism of piezoelectric generation is based on stable alignment of dipoles on mechanical force and its ability to recover from deformation. Therefore, if the orientation of dipoles is large, the polymer attains a higher dipole moment, leading to high piezoelectricity.²¹ Moreover, the piezoelectricity of polyvinylidene fluoride (PVDF), offered by the β phase crystalline structure polyacrylonitrile (PAN) also possesses piezo character due to the cyano (–CN) group in repeating units like (–F) in PVDF.²² Further, this amorphous polymer (vinyl-type) has two conformations, namely planar zig-zag and 3¹-helical structure. Among them, the zigzag conformation has the highest dipole moment because of the transform (TTTT) structure. Despite the high dipole moment produced by –CN groups, PAN is said to possess low piezoelectricity. However, the piezoelectric properties may be improved through the introduction of charge carriers or by better aligning the –CN groups.²³ In addition, utilizing the electrospinning technique for preparing PAN membranes can improve the zig-zag conformation and have stronger piezoelectric characteristics. Additionally, nanofibers can provide advantages such as flexibility, lightweightness, and improved transport of ions.²⁴ According to the reported literature, MXene, a novel 2D material, is of great interest in various fields such as catalysis, energy storage, strain sensors, EMI shielding, etc.^25–31 Owing to its unique layered structure and functionalization of the surface, it can provide superiority in terms of enhancing dipole orientation for harvesting as well as electrochemical properties by improving storage performance. Moreover, high surface area, good aqueous dispersion, and electrical conductivity are added benefits.³²

In this work, we have explored the PMX nanofibers as piezoelectric energy harvesters, PS, and electrodes in a self-charging energy storage device. First, the electrospun PMX nanofibers with varying MXene from 1 to 5 wt% were tested for energy harvesting. The PNG with PMX-3% showed high output voltage, current, and power density of 9.9 V, 1124 nA, and 9.6 μW cm⁻². Interestingly, we used the PMX-3% itself as electrodes in nanofiber form after the carbonization process. The symmetric nanofiber SCPS was fabricated with C-PMX-3% electrodes, PMX-3% as PS, and PVA/KOH gel electrolyte. Further, the influence of PS in improving the electrochemical performance and self-charging ability upon mechanical pressure due to the piezo-electrochemical conversion was studied and confirmed. Thus, a new multifunctional PMX nanofiber appeals as a potential candidate for self-charging applications.

2. Experimental section

2.1. Materials

Polyacrylonitrile powder (average M_W 150 000, Sigma-Aldrich), N,N-dimethylformamide (DMF, Sigma-Aldrich, 99%), hydrochloric acid (HCl; Honeywell, >37%), and ethanol (Sigma-Aldrich, 99%) were used. Lithium fluoride (LiF; 99.98%), DI water, and titanium aluminum carbide (Ti₃AlC₂ (MAX phase)) were purchased from Jilin 11 technology Co., Ltd. Poly(vinyl alcohol) (PVA, Alfa-Aesar, 99%), potassium hydroxide pellets (KOH, Honeywell, >85%), nickel foam, and carbon felt were also used. All the chemicals purchased were of analytical grade and were further used without any purification.

2.2. Synthesis of MXene

MXene (Ti₃C₂T_x) was obtained by selectively etching the aluminum (Al) layer from Ti₃AlC₂ using a blend of LiF and HCl. Initially, HCl (5 mL, 12 M) was added to the Teflon bottle containing 5 mL of deionized water. Gradually, LiF was added to the solution and stirred at 40 °C to ensure proper mixing. Next, the MAX phase was carefully introduced into the etching solution with magnetic stirring (360 rpm), and the etching reaction remained for 24 h at 40 °C. In the following step, the reaction product was centrifuged (3500 rpm for 5 minutes for each round) and washed with deionized water until its pH reached 7. Finally, an ice bath sonication of 30 min with subsequent centrifugation at 5000 rpm for 20 min yielded delaminated Ti₃C₂T_x.

2.3. Synthesis of PMX nanofiber separator

First, the electrospinning solutions at a concentration of 8 wt% of PAN were obtained by dissolving in DMF under vigorous stirring at 70 °C for 2 h. Then, various contents of MXene 1, 2, 3, 4, and 5 wt% were added to the PAN solution at 50 °C for 12 h. During the electrospinning procedure, the as-prepared PMX composite solution was loaded into a plastic syringe with a flow rate of 0.5 mL h⁻¹, applied voltage of 15 kV, and humidity (<35%). The PMX nanofibers were collected on aluminum foil, kept at a distance of 15 cm from the needle. For comparative purposes, a pure PAN nanofiber without the addition of MXene was prepared by the same procedure. Finally, the prepared nanofibers were termed as PAN, PMX-1%, PMX-2%, PMX-3%, PMX-4%, and PMX-5%.

2.4. Fabrication of piezoelectric nanogenerator

PAN and PMX composite nanofibers cut to 4 cm × 4 cm dimensions were placed between two aluminum foil electrodes that constituted the top and bottom electrodes. To connect the electrode externally, copper wires were used. During operation, adhesive tapes at the corners prevent air gaps from compromising the integrity of the sandwiched structure.

2.5. Synthesis of C-PMX electrodes

The C-PMX electrodes were obtained by stabilization followed by a carbonization process. Initially, the PMX nanofiber generated by electrospinning was stabilized in an oxygen atmosphere at 270 °C for 2 h. Henceforth, for carbonization, the PMX nanofiber was heated to 850 °C at a heating rate of 5 °C min⁻¹ under an argon atmosphere and kept at 850 °C for 2 h.

2.6. Preparation of gel electrolyte

First, 1 g PVA was dissolved in water and stirred for 4 h at 80 °C. After complete dissolution of PVA, 1 g of KOH was added, and the resulting solution was continuously stirred until formation of a homogeneous viscous solution.

2.7. Assembly of piezo-supercapacitor

The piezo-supercapacitor device was integrated in a glove box under vacuum. The symmetric self-charging supercapacitor is composed of three major components: an anode (C-PMX), a piezo-separator (PMX), and a cathode (C-PMX) sealed in a CR2032-coin-type cell filled with a PVA/KOH gel electrolyte. The detailed fabrication of SCPS with nickel foam (NF) and carbon felt (CF) is given in the SI.

2.8. Characterization

The structural morphology of the MXene and PMX nanofibers was studied by using field emission-scanning electron microscopy (FE-SEM, JEOL JSM-6700F) and field emission transmission electron microscopy (FE-TEM) imaging (Instrumentation Center, National Taiwan University). The surface potential was observed using a Kelvin probe force microscope (KPFM Bruker-Dimension Icon). Crystallography structure and changing phase polarization of MXene, PAN, and PMX were interpreted using the X-ray diffraction (XRD, PANalytical diffractometer (X’Pert3 Powder)) technique. Furthermore, Fourier transform infrared spectroscopy (FTIR, PerkinElmer, UK) was used. The capacitance was collected using a WAYNE KERR precision impedance analyser, 6420. The piezoelectric coefficients of the electrospun nanofibrous membranes were analyzed with a wide-range d₃₃ meter (d₃₃, YE2730, APC International Ltd, Mackeyville, PA, USA). A tensile test QC 508 (Cometech Testing Machines Co., Ltd, Taiwan) was conducted to determine the mechanical properties. Electrical performances related to voltage were studied through an oscilloscope. Short-circuit current and charge–discharge for SCPS-CF and SCPS-NF were studied with Keithley (2634B).

2.9. Electrochemical analysis

An Autolab PGSTAT204 electrochemical workstation was utilized to measure the energy storage and self-charging properties of SCPS. The cyclic voltammetry (CV) was measured in a potential window of 0–0.6 V at different scan rates. Galvanostatic charge–discharge (GCD) curves were recorded in the same potential range at different current densities. Further, the electrochemical impedance spectra (EIS) were carried out to understand the reaction kinetics and storage performance. The piezo-electrochemical behavior was confirmed by measuring CV and GCD with mechanical force (using an electrochemical workstation).

3. Results and discussion

3.1. Morphological and structural characterization of prepared MXene and electrospun PMX nanofibers

The schematic illustration in Fig. 1a depicts the process steps involved in the preparation of MXene. The cross-sectional and top-view morphology of the synthesized MXene is depicted in Fig. 1b and Fig. S1a, respectively. After HF treatment, the formation of a layered structure was attained due to the etching of Al layers from the MAX phase. The HR-TEM image (Fig. 1c) further confirmed the layer-to-layer structure of Ti₃C₂T_x. From the energy-dispersive X-ray analysis (EDAX) pattern of the MXene (Fig. S1b), the removal of Al was confirmed while verifying the presence of fluorine (F) and oxygen (O). This suggests that the exfoliated nanosheets may have surface terminations with F, OH, or O groups. As shown in Fig. 1d, the XRD patterns of the MAX phase show the diffraction peak of Al at 2θ = 39°, which disappeared in the MXene, showing the successful etching of Al layers. Interestingly, a backward shift in the (002) peak from 2θ = 9.5° to 2θ = 6.5° was attributed to the increased separation between the layers. Hence, the crystalline framework of the MAX phase vanishes, and the formation of a new layered structure is confirmed.

Fig. 1 (a) Schematic illustration of synthesis of MXene from MAX phase. (b) FE-SEM of prepared MXene (cross-section view). (c) FE-TEM surface morphology of MXene. (d) XRD patterns of MXene and its MAX phase.

In order to explore the effect of MXene concentration on fiber morphology, a spinning dope of different amounts (1–5 wt%) was prepared. The FE-SEM images of pure PAN and PMX-3% are shown in Fig. 2a and b with their morphology and diameter distribution histogram (inset), while composite fibers are shown in Fig. S2b–e. The electrospun nanofibers showed a continuous fiber with a uniform and smooth surface. PAN nanofibers had a mean diameter of 692 ± 4 nm, and the EDAX spectra in Fig. S2a confirmed the distribution of elements C and N in PAN nanofibers. After the addition of MXene, the fiber diameter decreased to 402 ± 5 nm, 369 ± 20 nm, 346 ± 2 nm, 435 ± 5 nm, and 493 ± 25 nm for PMX-1%, PMX-2%, PMX-3%, PMX-4%, and PMX-5%, respectively. The reduction in diameter is attributed to the improved static electric field force and solution conductivity. However, at 4 wt% and 5 wt% of MXene loading, the diameter becomes wider and bead formation occurs due to the increased viscidity of the polymer solution and aggregation of MXene. The magnified FE-SEM and EDAX image in Fig. 2c shows that PMX-3% contains Ti, C, and N elements, confirming the successful incorporation of MXene. Furthermore, XRD patterns of PAN and PMX composite fibers with various loadings of MXene depicted in Fig. 2d showed a distinctive peak at 2θ = 16.7° associated with the (100) crystal plane of PAN.³³ However, after the addition of MXene, the diffraction peak corresponding to the (100) plane shifted to the right with a maximum of 2θ = 17.2°. The right shift indicates the higher planar zig-zag conformation that arises due to the orientation of –CN groups of PAN that can induce high piezo-properties. In addition, the characteristic (002) plane appeared at 2θ = 6.2° for the composite nanofibers, confirming the successful incorporation of MXene into the PAN nanofibers. In the FT-IR spectra in Fig. 2e, the characteristic peaks at 2937 cm⁻¹ and 1452 cm⁻¹ correspond to the stretching and flexural vibrations of methylene (CH₂) of PAN. The distinctive peak at 2243 cm⁻¹ was associated with nitrile-based (–CN) stretching. In addition, slight shifts in the –OH and C=O peaks were observed, indicating interactions between the MXene and PAN. As can be seen in Fig. 2f, the peak at 1250 cm⁻¹ occurs due to the combination of wagging and rocking modes of methine (CH) methylene groups, while the methylene (CH₂) paired with (CN) and methine (CH) groups produces a peak at 1230 cm⁻¹ corresponding to the twisting mode. The following equation estimates the planar zigzag content (ϕ):

Fig. 2 FE-SEM images and average diameter distribution (inset) of (a) PAN and (b) PMX-3% nanofibers. (c) Magnified FE-SEM and EDS mapping of PMX-3% showing the distribution of Ti, C, and N. (d) XRD of PAN and PMX nanofibers. (e) FT-IR spectra of PAN and its composite fibers in the frequency range of 400–4000 cm⁻¹. (f) FT-IR spectra of PAN and its composite fibers in the frequency range of 1200–1300 cm⁻¹. (g) Stress–strain curves for assessing the mechanical performance of PAN and PMX nanofibers. (h) Frequency-dependent dielectric constant of prepared nanofibers. (i) Piezoelectric coefficient of PAN and PMX with different MXene content.

S ₁₂₅₀ and S₁₂₃₀ were the peak areas at 1250 cm⁻¹ and 1230 cm⁻¹ corresponding to planar zig-zag conformation and 3¹-helical conformation, respectively.³⁴ The different loadings of MXene varied the zig-zag content that enhanced the piezoelectric properties of the fibers. As shown in Fig. S3 (bar graph), the ϕ of PMX-1%, PMX-2%, PMX-3%, PMX-4%, and PMX-5% was 43.4%, 48.1%, 59.3%, 54.6%, and 51.7%, respectively, which was higher than PAN nanofibers (ϕ = 38.2%). Thus, we can conclude that PMX-3% has a higher ϕ content that can yield higher piezoelectric properties. Nevertheless, at 4% and 5% doping, a decrease in zig-zag content indicated the agglomeration of MXene. Furthermore, the zig-zag orientation offers high orientation (all transform TTTT) in the polymer chain, providing better mechanical and dielectric properties. Notably, from Fig. 2g, the MXene incorporation into the PAN nanofibers led to better stress and strain tolerance. Specifically, PMX-3% showed a three-fold increase in stress and doubled strain values because of the establishment of strong interaction between the termination groups of MXene and the nitrile group of PAN. Conversely, the excessive doping had no contribution to improvement due to particle agglomeration by MXene in the PAN matrix. This result is in good agreement with the FT-IR and XRD data. The dielectric constant data of PAN, PMX-1%, PMX-2%, PMX-3%, PMX-4%, and PMX-5% in the frequency range of 10³–10⁶ Hz are delineated in Fig. 2h. In this regard, pure PAN fibers had a lower dielectric constant value than that of the composite fibers from the analysis. Moreover, MXene addition in various wt% had a progressive impact with increasing concentration up to 3 wt% due to the high interfacial polarization. However, at higher doping, the dielectric constant value decreased. Despite the decrease, the value was still higher compared to PAN nanofibers. The highest dielectric constant of 10.7 (five times higher) was provided by PMX-3%, while PAN's was 1.8. According to the percolation theory, after a particular concentration referred to as ‘percolation threshold,’ agglomeration of particles inhibits the further interfacial polarization, resulting in the diminution of its dielectric nature. Notably, the piezoelectric coefficient (d₃₃) is a key indicator of material performance in piezoelectric applications. As shown in Fig. 2i, the d₃₃ of PMX-3% attains a maximum value of 43.9 pC N⁻¹, ascribed to the enhanced dielectric constant. It is worth noting that the presence of MXene up to 3 wt% has an increasing effect on the zig-zag phase enhancement as well as the dielectric nature of the prepared nanofibers. However, at high concentrations, agglomeration suppresses the phase transformation and reaching of the percolation limit, which is in good agreement with dielectric constant results.

3.2. Surface potential analysis of PMX nanofibers

KPFM is a reliable technique for the quantification of surface charges that can be related to the charge polarization in piezoelectric devices during performance. The schematic representation of KPFM measurements on PMX nanofibers is illustrated in Fig. 3a. A surface topographical image of PAN and PMX-3% with their respective charge potential distribution is shown in Fig. 3b–e, respectively. Moreover, the surface potential distributions of PMX-1%, PMX-2%, PMX-4%, and PMX-5% are provided side-by-side, enabling comparison in Fig. S4a–d, respectively. It can clearly be seen that the surface potential shifts more towards a negative value upon MXene addition. Notably, PMX-3% had the highest color contrast in surface potential measurements and even distribution of MXene. Further KPFM signal profiles were analyzed to understand the evolution of the KFPM contrast. As in Fig. 3f, the negative surface potential of PAN fibers (−1.7 V) is due to the electronegative –CN groups. Evidently, the increasing loadings of MXene increased the negative surface potential, affirming that MXene influences the polarization in PAN molecules. PMX-3% showed the highest negative potential of −4.6 V by inducing a strong and localized alignment of dipoles.

Fig. 3 (a) Schematic representation of KPFM measurements on PMX nanofibers. (b) Topographic image of PAN nanofibers. (c) Surface potential image of PAN nanofibers. (d) Topographic image of PMX-3% nanofibers. (e) Surface potential image of PMX-3% nanofibers. (f) Variation of surface potential in line graph scan results.

3.3. Mechanical energy harvesting properties of PMX nanofiber-based nanogenerators

The detailed working mechanism of the PNG is illustrated in Fig. S5. The energy harvesting from piezo materials can be performed in two operation modes, namely 33-mode and 31-mode, as seen in Fig. S6. Herein, the ‘3’ direction represents the polar axis, while ‘1’ represents right angles. In the case of 33-mode, the applied stress and voltage obtained are in the same direction. However, 31-mode produces a voltage in a perpendicular direction while the stress acts in an axial direction. Furthermore, the operation mode significantly alters the output voltage: 33-mode produces higher voltage, whereas 31-mode offers high current.³⁵ Hence, in this work, we have utilized 33-mode operation and investigated the piezoelectric performance of the devices fabricated with PAN, PMX-1%, PMX-2%, PMX-3%, PMX-4%, and PMX-5% nanofibers by employing a vertical compression and release. Mechanical deformation on the piezoelectric devices generated a piezoelectric potential due to the specific molecular dipole alignment. The applied force alters the distance and direction of the –CN group and hydrogen, resulting in an altered polarization strength as well as orientation of dipoles. Consequently, the open-circuit voltage increases due to an ion concentration gradient. The successive press and release caused the shift of the center of mass by aligning molecular dipoles, resulting in the generation of positive and negative piezoelectric voltage. Fig. 4a and b show the open-circuit voltage and short-circuit current of devices made using PAN and PMX piezo fibers with varying MXene content. It can be seen that the output voltage and current gradually increased up to 3 wt% of MXene, which then decreased on further MXene addition. The decreased performance at higher concentrations of MXene is due to the agglomeration forming conductive networks that can neutralize the charge density. A maximum open-circuit voltage and short-circuit current of 9.9 V and 1124 nA were recorded for PMX-3%, respectively, due to its highest zig-zag content. To verify that the electrical output originated from the piezoelectric effect, polarity-switching measurements of the PNGs were conducted. As illustrated in Fig. 4c, the PNG exhibited a clear periodic alternation of positive and negative output responses during repeated compression and release cycles. The reversal of the voltage signal upon switching the electrode connections is attributed to the polarization reversal of the dipoles. These findings confirm that the PNG's output is primarily driven by piezoelectric effects rather than triboelectric contributions. The rectified output voltage displayed in Fig. 4d represents all positive DC signals with a magnitude decline due to the rectifying diodes in the circuit. As shown in Fig. S7, the rectification circuit enables the storage of harvested energy. Henceforth, we illustrated the charging of various commercial capacitors (1, 3.3, and 22 μF) using a PMX-3% device through a bridge rectifier (Fig. 4e). Moreover, the external load resistance-dependent output voltage was obtained by increasing the load from 1 to 8 MΩ. As shown in Fig. 4f, the voltage increased with increasing resistance, and a maximum power density of 9.6 μW cm⁻² was obtained at a 6 MΩ load. The piezo output of our device was compared with previous reports in Table 1. A gradual increase in the output voltage was observed with increasing frequency applied, as shown in Fig. 4g. As the frequency increases, the internal impedance of PMX decreases at the strained state. At low frequency, the electron flow in the external circuit is slow, resulting in easy neutralization of accumulated charges that produce small voltage. However, at high frequency (>3 Hz), PNG generated high voltage due to high electron flow that did not neutralize quickly. In order to test the long-term use, a stability test was performed. It was observed that the PMX-3% piezo device performance was retained without any decreased voltage for 1000 s (Fig. 4h).

Fig. 4 (a) Voltage output of devices fabricated with PAN, PMX-1%, PMX-2%, PMX-3%, PMX-4%, and PMX-5%. (b) Short-circuit current characteristics of PAN and its composite devices. (c) Forward–reverse switching polarity test of the PMX-3% device. (d) Rectified voltage of PMX-3% with bridge rectifiers. (e) Charging of various capacitors by the PMX-3% device. (f) Output voltage and power density as a function of resistors from 1 to 8 MΩ. (g) Voltage of PMX-3% device on different applied frequency. (h) Stability performance of the PMX-3% device.

Table 1 Comparison of output performance of piezo device PMX-3% with previous reports

Nanogenerator	Poling Treatment	Voltage (V)	Current	Power density	Ref.
PAN	Electrospinning	2	1.2 μA	0.91 μW	22
PAN/BaTiO3/Eu^3+	Electrospinning	6	45 nA	—	41
PAN/BaTiO3	Electrospinning	9.3	0.53 mA	—	23
ZnO/PAN	Electrospinning	4.3	1.8 μA	9.5 mW m^−2	34
PAN/TMAB	Electrospinning	2.56	0.61 μA	0.19 μW	42
PAN/CuO	Electrospinning	5	0.172 μA	0.215 μW cm^−2	43
PVDF/PAN	Electrospinning	1.3	0.32 μA	—	44
PANI/PAN	Electrospinning	0.5	—	—	45
PAN/FeCl3 · 6H2O/ILs	Electrospinning	5.2	1.5 μA	—	46
PMX	Electrospinning	9.9	1124 nA (1.124 μA)	9.6 μw cm ^−2	This Work

---

3.4. PMX nanofiber-based SCPS

Fiber-based electrodes for energy storage applications are highly demanded due to their light weight and flexibility and can offer high surface area.^36–38 In addition, the nanofibrous structure can offer better ion transfer efficiency that can improve the energy density of supercapacitors. Fig. 5a illustrates the electrospinning of PMX piezofibers and the preparation process of C-PMX electrodes. The structure of the symmetric piezo-supercapacitor is fabricated with C-PMX electrodes separated by PMX PS and PVA/KOH gel electrolyte. After preparation, the PMX-3% nanofiber was subjected to stabilization at 270 °C to promote cross-linking by cyclization and dehydrogenation. Subsequently, carbonization of the stabilized sample at 850 °C under an inert Ar atmosphere provided a structural transformation to a graphitic phase, resulting in improved conductivity and mechanical strength. Further, from the SEM images of the stabilized and carbonized sample in Fig. 5b and c (inset—average diameter distribution), it could be inferred that the average fiber diameter decreased from 338 ± 3 nm to 320 ± 2 nm due to the mass loss that occurred as a result of pyrolysis treatment. The bar graph in Fig. 5d reveals the decreasing average fiber diameters of as-spun PMX-3%, stabilized, and carbonized PMX-3%, respectively. In addition, in the XRD spectra recorded in Fig. 5e, PMX-3% showed a diffraction peak at 16.8° related to the (100) plane. However, upon stabilization and carbonization, this (100) plane disappeared, and a new peak at 24° related to (002) arose, indicating the formation of a layered graphitic structure. This confirms the destruction of the PAN structure and the formation of a new material. Photographic images of stabilized and carbonized PMX-3% are shown in Fig. 5f and g, respectively. After the carbonization process, the conductivity of the PMX-3% nanofiber was tested using a multimeter (Fig. S8) and was directly used as positive and negative electrodes.

Fig. 5 (a) Illustration of the preparation process of C-PMX electrodes from electrospun PMX piezofibers and fabrication of a symmetric piezo-supercapacitor with PMX PS and C-PMX electrodes. FE-SEM and average diameter distribution (inset) of PMX-3% nanofibers. (b) Stabilized. (c) Carbonized. (d) Comparison of average diameters of PAN and PMX-3% (stabilized and carbonized). (e) XRD patterns of PAN, stabilized and carbonized PMX-3%. Photographic images of PMX-3%: (f) stabilized and (g) carbonized.

Generally, a supercapacitor device consists of two electrodes, an electrolyte and a separator. The electrodes undergo adsorption/desorption or redox reactions based on the nature of the electrode material used and store charges. While the electrolyte supplies ions for electrochemical reaction and the separator prevents a short circuit between the electrodes. After the supercapacitor is fully charged, self-discharge can happen due to the ion migration from high to low concentration resulting in loss of charges and reduce stored energy. Interestingly, when the separator used is replaced by a piezoelectric nanofiber membrane, the piezoelectric voltage generated drives the electrolyte ions towards the electrode surface resulting in the storage of electrochemical energy. The PS employed has a polarized charge on its surface that prevents the charge diffusion from the electrode surface and prolongs the charge storing as in Fig. S9 which cannot be attained with normal separators. First, the influence of PMX-3% (PS) on the energy storage performance was studied by assembling PMX-3% (PS) and C-PMX-3% as positive and negative electrodes in the CR2032 coin cell with PVA/KOH gel electrolyte. For comparative purposes, another device without PMX-3% was fabricated using the same procedure. Similarly, the stabilized sample was also tested for a clearer understanding of the importance of the carbonization step for storage performance. Fig. 6a and b show the CV profiles of the device fabricated with and without PS at different scan rates (10–100 mV s⁻¹) in a potential window of 0–0.6 V. The CV curve showed a quasi-rectangular shape without any redox peaks, revealing electrical double layer capacitance (EDLC) behavior of C-PMX-3% composite fibers. It can be clearly seen that the integrated area of the curve for the PMX-3%-incorporated device is increased due to the improved electrolyte ion reaction kinetics at the electrode/electrolyte interface due to the polarization effect of the PMX-3%. Further, the increased scan rate showed increased curve area without significant distortion, indicating the good electrochemical behavior of the fabricated device. Similarly, the CV profiles of devices fabricated with stabilized samples in Fig. S10a and b showed improved integrated area with increasing scan rate, but the current generated was small compared to carbonized sample. In addition, the influence of PS was visible, but its contribution to device performance is negligible. The GCD curves at different current densities (1–5 A g⁻¹) of the PMX-3% incorporated device are illustrated in Fig. 6c. The shape of an isosceles triangle was observed for the charging and discharging curve, demonstrating dominating EDLC behavior and stable electrochemical properties of SCPS fabricated with PMX-3% and C-PMX-3% electrodes. From the discharge curve at 1 A g⁻¹, the specific capacitance of the device was calculated to be 100.5 F g⁻¹ using eqn (S1). The superiority of SCPS with PMX-3% performance was compared with previously reported works as shown in Table 2. Further, the SCPS delivered an energy density of 5 Wh kg⁻¹ at a power density of 300 W kg⁻¹ (calculated using eqn (S2) and (S3)). The EIS spectra in Fig. 6d show the Nyquist plots of SCPS with and without the PMX-3% separator. At the electrode–electrolyte interface, ions migrate at a resistance determined by the diameter of the semicircle. Therefore, a large semi-circle diameter at the high-frequency region could be observed for the device fabricated without PMX-3%, suggesting greater charge transfer resistance. Meanwhile, a device with PMX-3% reveals faster ion diffusion due to the decreased fiber diameter and increased conductivity of PMX-3% by the carbonization process. However, the low-frequency region reveals the slow transport kinetics via the slope of a straight line due to the ions that cannot diffuse through MXene but only around them. In Fig. S11, the Nyquist plot shows a nearly linear profile in both cases, indicating dominant capacitive behavior with minimal charge-transfer resistance, a typical feature of electric double-layer capacitors (EDLCs). Despite this small change in internal resistance, the device retained a low resistance and maintained efficient ion transport, confirming stable electrochemical performance. Moreover, the inclusion of PMX-3% influenced the capacitance retention over cycles. The SCPS device with PS showed an excellent capacitance retention of 92% even after 5000 cycles at 1 A g⁻¹ (Fig. 6e) due to the synergetic effect of the conversion of strain to charge along with improved ion distribution. However, the retention ability without PS had a faster decline of 85% after 5000 cycles due to a lack of the regeneration mechanism and storage limited to electrochemical cycling. Thus, we can conclude that the PMX-3% PS has an upper hand in improving storage performance compared to regular separators.

Fig. 6 CV profiles recorded with different scan rates from 10 to 100 mV s⁻¹ for SCPS fabricated with carbonized sample (a) without PMX-3% piezo-separator. (b) with PMX-3% PS. (c) GCD with varying applied currents from 1-5 A g⁻¹ for SCPS with PMX-3%. (d) Comparison of EIS in the form of a Nyquist plot for SCPS with and without PMX-3%. (e) Capacitance retention of SCPS device over 5000 cycles at 1 A g⁻¹. Piezo-electrochemical response. (f) CV profiles recorded by applying different pressures. (g) Comparison of CV recorded with and without tap and pressure. Charging of the piezo-supercapacitor device recorded for 12 s by (h) compression and (i) tapping.

Table 2 Comparison of output performance of SCPS with previous reports

Electrode	Electrolyte	Separator	Specific capacitance	Ref.
Graphene sheet	TEABF4	Porous PVDF	28.46 F g^−1	47
MnO2-rGO	PVA/H3PO4	PVDF-rGO-ZnO	7.6 F g^−1	48
Ag nanowire/NiO	PVA/KOH	PVDF-TrFE	3.47 mF cm^−2	49
MoSe2	PVDF-co-HFP/TEABF4	PVDF/NaNbO3	18.93 mF cm^−2	50
MnO2	PVA/H3PO4	PVDF-ZnO	455 mF g^−1	51
C-PMX	PVA/KOH	PAN/MXene	100.5 F g ^−1 or 201 mF cm ^−2	This Work

---

To understand the piezo-electrochemical process in detail, the SCPS was subjected to mechanical force such as tapping or compression. The CV profile of SCPS with various applied pressures, i.e., low, medium, and high, showing a variation in current concerning varying pressure due to the piezo ionic movement, is shown in Fig. 6f. This current variation arises due to the different concentrations of electrolyte ions at applied pressures. Interestingly, a current spike with positive and negative polarity is observed in the forward and reverse cycle, respectively. From this, we can understand that the SCPS possesses high sensitivity to imparting pressure. The emergence of the current spikes and the additional current can be attributed to the piezo-electrochemical process. Fig. 6g illustrates the variation in the CV profile obtained on subjecting the device with and without a tap and pressure. The piezoelectric properties of SCPS were further confirmed by a charging process with compression and tapping. Herein, the compression force led to the charging of the device from 0 to 47 mV in 12 s, as shown in Fig. 6h. However, the tapping produced a charge from 0 to 23 mV in 12 s (Fig. 6i). The detailed self-charging mechanism through piezo-electrochemical conversion is illustrated in Fig. 7a. Initially, the SCPS device is at a discharged state (Fig. 7a(i)). Next, when the device experiences a mechanical force, PS becomes polarized (Fig. 7a(ii)). An internal electromagnetic field is generated as a result of polarization, creating a positive and negative potential difference at the top and bottom of the PS. This piezoelectric effect decides the movement of K⁺ and OH⁻ ions towards the respective electrodes (piezoelectric potential-driven ion migration). Hence, the distributed positive and negative ions at the electrode surface cause potential differences between the anode and cathode of the supercapacitor, causing the self-charging of the SCPS device (Fig. 7a(iii)). Once the external force is stopped, the potential difference starts to decline due to the falling back of electrolyte ions, and the whole system becomes neutral (Fig. 7a(iv)). Additionally, the practical applicability of PMX as PS and C-PMX as the electrode was tested by fabricating the piezo-supercapacitor device with CF and NF substrates. The devices were named SCPS-CF and SCPS-NF. The detailed fabrication process is provided in the SI. The carbon felt was chosen due to its porous nature, better electrical conductivity, cost-effectiveness, and corrosion resistance.³⁹ Also, the fiber structure of CF can offer flexibility, high surface area, and improved mechanical as well as electrochemical properties. The excellent mechanical and flexible properties of SCPS-CF are illustrated in Fig. S12(a). Interestingly, fiber-based self-charging devices can be achieved because they are both fibers, and all fiber-based self-charging devices can be attained, as the PMX-3% and C-PMX are fibers as well, along with the substrate. The self-charging studies for SCPS-CF showed a potential of 250 μV (Fig. 7b). Notably, the device was charged from 55 μV to 250 μV in just 0.2 s, while the discharge time took about 1.5 s to decline from 250 μV to 48 μV back. Compared to periodic bending (Fig. 7c), continuous tapping bending (Fig. 7d) produced more output. The self-charging response of SCPS-CF with a slow tap and fast tap, along with compression and tapping, is demonstrated in Video S1. The change in voltage of the device under pressure/compression indicates the piezoelectric effect due to the deformation of PMX-3%. On the other hand, NF also has a porous nature, good electrical conductivity, and is cheap.⁴⁰ In addition, it provides good mechanical stability under controlled conditions such as optimum KOH concentration, and the potential window can prevent the possibility of substrate corrosion during electrochemical processes. SCPS-NF also shows good flexibility, as shown in Fig. S12(b). SCPS-NF produced a potential of 4 mV (Fig. 7e). However, in SCPS-NF, varying tapping and compression can generate from 3 to 7 mV, as demonstrated in video 2. Strikingly, the switching of the polarity was reflected in the charge discharge of SCPS-NF in Fig. 7f and g. On account of the signal's reversibility, the piezo-electrochemical effect can be confirmed. Also, the rectified signal obtained gives an upper hand in terms of rectification-free charge–discharge. Thus, we can conclude that the piezo behavior of PMX-3% influences and enhances the performance of the SCPS device. Therefore, PMX-3% with piezoelectric properties can suppress the discharge and improve the performance of self-charging supercapacitors.

Fig. 7 (a) Mechanism of charge storage in SCPS: (i) SCPS at initial state without compressive force; (ii) compressive force generating piezo charges at PMX driving electrolyte ions to electrodes; (iii) equilibrium state between piezo potential and electrochemical reaction; (iv) return of the device to its initial state after completion of one charging cycle. Self-charging response of SCPS-CF (b) subjected to compression and release, (c) upon periodic bending, and (d) under continuous tapping. (e) Self-charging performance of SCPS-NF device subjected to compression and release. Rectified charging of SCPS-NF: (f) forward and (g) reverse.

4. Conclusion

In summary, the experimental findings conclude that the PMX nanofiber prepared by electrospinning can be utilized as both an electrode and a PS for self-charging devices with different substrates. The carbonization step helped in enhancing the conductivity of PMX-3% for its potential use in SCPS. The incorporation of MXene helped in enhancing the piezo properties of PAN by increasing the zig-zag content, confirmed by various studies. The output performance of the PNG was superior compared to previous reports based on PAN. Employing PMX-3% as a PS in SCPS fabricated with C-PMX suppressed the energy loss and also delivered superior performance metrics in terms of efficient conversion of harvested energy. Moreover, the key findings of this work can lay the foundation for the development of rectification-free integrated devices for sustainable future electronics.

Author contributions

Jayashree Chandrasekar: conceptualization, methodology, investigation, writing – original draft. Manikandan Venkatesan: conceptualization, methodology, formal analysis, writing – review. Chen-Wei Fan, Hao-Yuna Chen: software, formal analysis. Yung-Chi Hsu, Wei-Wen Chen: methodology, formal analysis. Ming-An Chung: methodology. Mei-Wan Chung: methodology. Wen-Ya Lee: methodology, investigation. Ja-Hon Lin: investigation. Ye Zhou: investigation, formal analysis, writing – review and editing. Chi-Ching Kuo: supervision, investigation, formal analysis, writing – review and editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5mh01510c.

Acknowledgements

This work was financially supported by the “Advanced Research Center for Green Materials Science and Technology” from the Featured Area Research Center Program within the framework of the Higher Education Sprout Project by the Ministry of Education (113L9006). We also acknowledge the financial support provided by the National Science and Technology Council (Contracts: NSTC 112-2221-E-027-003-MY3) and the NTUT-SZU Joint Research Program (Grant No. NTUT-SZU-114-01).

References

1. Z. L. Wang, Adv. Mater., 2012, 24, 280–285.
2. V. Z. Gjorgievski, N. Markovska, B. V. Mathiesen and N. Duić, Smart Energy, 2022, 8, 100091.
3. F. Chien, L. Huang and W. Zhao, J. Innov. Knowl., 2023, 8, 100298.
4. M. A. Hannan, S. Mutashar, S. A. Samad and A. Hussain, Biomed. Eng. Online, 2014, 13, 79.
5. R. Yang, Y. Qin, L. Dai and Z. L. Wang, Nat. Nanotechnol., 2009, 4, 34–39.
6. L. Gu, J. Liu, N. Cui, Q. Xu, T. Du, L. Zhang, Z. Wang, C. Long and Y. Qin, Nat. Commun., 2020, 11, 1030.
7. L. M. dos Santos, F. L. Santos, F. Scinocca and G. P. Villibor, Energy Rep., 2025, 13, 455–464.
8. Y. Zheng, H. Yuan, W. Feng, Y. Qu and Y. Luo, J. Mec. Theor. Appl., 2025, 111, 105554.
9. C. Cao, P. Zhou, W. Qin, J. Wang, M. Liu, P. Wang, T. Zhang and Y. Qi, Appl. Surf. Sci., 2025, 685, 162028.
10. V. Thulasi, P. Lakshmi, E. James Jebaseelan Samuel and S. Mohana Roopan, Mater. Sci. Eng. B, 2025, 313, 117915.
11. P. Sengupta, A. Ghosal, S. Haldar and R. Ray, Chem. Eng. J., 2025, 503, 158486.
12. Q. Zhou, C. Zhu, H. Xue, L. Jiang and J. Wu, ACS Appl. Mater. Interfaces, 2024, 16, 35268–35278.
13. S. Selvam, S. Praveenkumar and J.-H. Yim, Chem. Eng. J., 2024, 482, 148929.
14. N. U. H. L. Ali, P. Pazhamalai, A. Sathyaseelan, T. D. Dongale and S.-J. Kim, Appl. Energy, 2024, 376, 124265.
15. C. Zhao, J. Xu, J. Tao, C. Xiao, X. Jin, W. Wang, X. Liu, J. Chen and Z. Zhu, Nano Energy, 2023, 116, 108812.
16. A. De Adhikari, S. Singh and I. Lahiri, J. Alloys Compd., 2023, 939, 168713.
17. R. Song, H. Jin, X. Li, L. Fei, Y. Zhao, H. Huang, H. Lai-Wa Chan, Y. Wang and Y. Chai, J. Mater. Chem. A, 2015, 3, 14963–14970.
18. D. Zhou, N. Wang, T. Yang, L. Wang, X. Cao and Z. L. Wang, Mater. Horiz., 2020, 7, 2158–2167.
19. M. Weng, J. Zhou, Y. Ye, H. Qiu, P. Zhou, Z. Luo and Q. Guo, J. Colloid Interface Sci., 2023, 647, 277–286.
20. H. Wang, Q. Zhou, B. Yao, H. Ma, M. Zhang, C. Li and G. Shi, Adv. Mater. Interfaces, 2018, 5, 1701547.
21. W. Peng, J. Chang, B. Wang, Z. Liu, G. Wang and S. Dong, Ceram. Int., 2024, 50, 6598–6605.
22. W. Wang, Y. Zheng, X. Jin, Y. Sun, B. Lu, H. Wang, J. Fang, H. Shao and T. Lin, Nano Energy, 2019, 56, 588–594.
23. L. Yuan, W. Fan, X. Yang, S. Ge, C. Xia, S. Y. Foong, R. K. Liew, S. Wang, Q. Van Le and S. S. Lam, Composites Commun., 2021, 25, 100680.
24. J. Fang, X. Wang and T. Lin, J. Mater. Chem., 2011, 21, 11088.
25. Z. Wang, Z. Cheng, C. Fang, X. Hou and L. Xie, Composites, Part A, 2020, 136, 105956.
26. M. Sang, S. Liu, W. Li, S. Wang, J. Li, J. Li, S. Xuan and X. Gong, Composites, Part A, 2022, 153, 106727.
27. Y. Yang, Y. Xu, Q. Li, Y. Zhang and H. Zhou, J. Mater. Chem. A, 2022, 10, 19444–19465.
28. A. Sfeir, C. E. Shuck, A. Fadel, M. Marinova, H. Vezin, J.-P. Dacquin, Y. Gogotsi, S. Royer and S. Laassiri, J. Am. Chem. Soc., 2024, 146, 20033–20044.
29. X. Li, Z. Huang, C. E. Shuck, G. Liang, Y. Gogotsi and C. Zhi, Nat. Rev. Chem., 2022, 6, 389–404.
30. P. A. Shinde, A. M. Patil, S. Lee, E. Jung and S. Chan Jun, J. Mater. Chem. A, 2022, 10, 1105–1149.
31. W. Ur Rehman, Z. Khan, F. Zahra, A. Laaskri, H. Khan, U. Farooq, M. Bajaj and I. Zaitsev, Sci. Rep., 2024, 14, 25107.
32. Z. Fu, N. Wang, D. Legut, C. Si, Q. Zhang, S. Du, T. C. Germann, J. S. Francisco and R. Zhang, Chem. Rev., 2019, 119, 11980–12031.
33. Y. Huang, R. Xue, L. An, Q. Shi, W. Zhang, Y. Wu and R. Liu, Smart Mater. Struct., 2023, 32, 095018.
34. Y. Sun, Y. Liu, Y. Zheng, Z. Li, J. Fan, L. Wang, X. Liu, J. Liu and W. Shou, ACS Appl. Mater. Interfaces, 2020, 12, 54936–54945.
35. N. Sezer and M. Koç, Nano Energy, 2021, 80, 105567.
36. J. Kim, J. G. Kang, J. Choi, P. V. Braun and S.-K. Kim, ACS Appl. Energy Mater., 2021, 4, 13716–13724.
37. P. Teymoory, J. Zhao and C. Shen, Micromachines, 2023, 14, 1249.
38. J. Liang, G. Zhu, C. Wang, Y. Wang, H. Zhu, Y. Hu, H. Lv, R. Chen, L. Ma, T. Chen, Z. Jin and J. Liu, Adv. Energy Mater., 2017, 7, 1601208.
39. J. Saupsor, J. Sangsawang, W. Kao-ian, F. Mahlendorf, A. A. Mohamad, R. Cheacharoen, S. Kheawhom and A. Somwangthanaroj, Sci. Rep., 2022, 12, 21156.
40. D. Wang, Y. Xu, X. Guo, Z. Fu, Z. Yang and W. Sun, J. Alloys Compd., 2021, 870, 159472.
41. Y. He, G. Fu, D. He, Q. Shi and Y. Chen, Mater. Lett., 2022, 314, 131888.
42. X. Li, Y. Li, Y. Li, J. Tan, J. Zhang, H. Zhang, J. Liang, T. Li, Y. Liu, H. Jiang and P. Li, ACS Appl. Mater. Interfaces, 2022, 14, 46789–46800.
43. S. Bairagi, A. Chowdhury, S. Banerjee, A. Thakre, A. Saini and S. W. Ali, J. Mater. Sci. Mater. Electron., 2022, 33, 13152–13165.
44. Z. Liu, G. Li, Q. Qin, L. Mi, G. Li, G. Zheng, C. Liu, Q. Li and X. Liu, Adv. Compos. Hybrid Mater., 2021, 4, 1215–1225.
45. K. Hu, J. Feng, Q. Hai, W. Jiang, Z. Lyu and N. Lv, Colloids Surf., A, 2022, 641, 128592.
46. Q. Shi, S. He, Y. He, Y. Wu and R. Liu, J. Appl. Polym. Sci., 2023, 140, e53824.
47. S. Sahoo, K. Krishnamoorthy, P. Pazhamalai, V. K. Mariappan, S. Manoharan and S.-J. Kim, J. Mater. Chem. A, 2019, 7, 21693–21703.
48. A. Rasheed, W. He, Y. Qian, H. Park and D. J. Kang, ACS Appl. Mater. Interfaces, 2020, 12, 20891–20900.
49. S. Qin, Q. Zhang, X. Yang, M. Liu, Q. Sun and Z. L. Wang, Adv. Energy Mater., 2018, 8, 1800069.
50. P. Pazhamalai, K. Krishnamoorthy, V. K. Mariappan, S. Sahoo, S. Manoharan and S. Kim, Adv. Mater. Interfaces, 2018, 5, 1800055.
51. A. Ramadoss, B. Saravanakumar, S. W. Lee, Y.-S. Kim, S. J. Kim and Z. L. Wang, ACS Nano, 2015, 9, 4337–4345.

---