The ternary nanocomposite g-C₃N₄/TiO₂@polyaniline embedded in PVDF based triboelectric nanogenerator for energy harvesting

Abstract

We report a materials-engineering strategy to enhance triboelectric nanogenerator (TENG) output by embedding a ternary g-C₃N₄/TiO₂@polyaniline (GTP) nanocomposite into PVDF tribopositive electrodes and pairing them with PTFE tribonegative layers of systematically varied composition from 30 to 60 wt%). XPS, TEM, FESEM/EDX, UV–Vis, and DSC confirm that TiO₂ nanoparticles are well distributed on g-C₃N₄ nanosheets and interconnected through a conducting PANI network, with evidence of interfacial Ti–O–C/Ti–N bonding, Ti³⁺ associated defect states, and band-tail absorption (Urbach energy, E_U = 0.31 eV), indicating localized electronic states that support charge storage/transfer.” In vertical contact–separation mode, the devices exhibit a composition-dependent trade-off: the 30 wt% PTFE configuration delivers the highest short-circuit current (Isc = 2.86 μA) and peak power (P_max = 10.3 μW), whereas the 60 wt% PTFE device yields the maximum open-circuit voltage (Voc = 4.56 V). These results demonstrate that voltage–current characteristics can be tuned through interfacial state engineering and controlled adjustment of the tribonegative layer composition. To confirm repeatability and stability under manual tapping, the best-performing 30 wt% and 60 wt% devices has been validated using an automated cycle-resolved statistical analysis.

Introduction

Triboelectric nanogenerators (TENGs), which transform mechanical energy such as that found in motion, vibration, and human activity into electrical energy, have become a particularly appealing option for low-power electronics due to their straightforward device architectures, wide range of materials, environmental friendliness, high instantaneous power density, and inexpensive manufacturing methods, and ability to deliver a significant electrical output appropriate for powering sensors and low-energy devices. The triboelectric nanogenerator (TENG), an effective device that can transform mechanical stimulation into electrical output, was initially developed by Z. L. Wang¹ and his team in 2012². The triboelectric nanogenerator works on triboelectrification and that is the combined effect of contact electrification and electrostatic induction. The process of bringing two distinct electron affinities of triboelectric materials into physical contact either by pressing, sliding, or separating is known as triboelectrification. Both electrodes separate when the external force is released, causing the potential to drop suddenly. Electrons move between the electrodes to counterbalance this, completing the half-cycle of the TENG’s AC output. This process repeatedly occurs and provide positive and negative output voltage and current. The current output of TENG device is called displacement current³. The output performance of TENG device critically depend on surface functionalization, interfacial charge trapping and transport, and the dielectric and mechanical properties of the contacting layers^4,5. TENG performance and stability can therefore be enhanced by designing materials that increase charge creation, inhibit recombination, and optimize charge extraction.

Most commonly utilized polymer materials include conducting polymers like polyaniline (PANI) and poly(3,4-ethylenedioxythiophene):polystyrene sulfonate (PEDOT:PSS), water-soluble/processible polymers like polyvinylpyrrolidone (PVP), and electroactive and dielectric polymers like poly(vinylidene fluoride) (PVDF), polytetrafluoroethylene (PTFE), and polydimethylsiloxane (PDMS)^6,7. Due to their inherent propensity for transferring or accept electrons, polymers like PTFE, PDMS, and PVDF hold crucial positions in the triboelectric series. Polytetrafluoroethylene (PTFE) is considered a highly tribo-negative material due to its remarkable chemical and thermal durability, low surface friction, and great electrical insulation. Haiyang Zou et al.⁸ stated that PTFE has a triboelectric charge density of −113 μCm⁻². Although poly (vinylidene fluoride) (PVDF) combines triboelectric and piezoelectric responses in a unique way, its low electrical conductivity prevents direct charge transport to external circuits^9,10. To get around this, PVDF is frequently combined with conductive pathways or charge-transfer agents (such as conductive fillers, interfacial electrodes, or dopants) to enable effective charge extraction^11,12. PVDF was selected as the host polymer for the ternary g-C₃N₄/TiO₂@PANI (GTP) filler because of its excellent film-forming ability, mechanical flexibility, and ferroelectric β-phase, which can enhance interfacial polarization and charge storage in composite triboelectric layers. Functional fillers provide a practical way for modifying microstructure, surface affinity, and dielectric constant, which in turn can modify charge generation, storage, and retention. In this regard, conductive or redox-active polymers have garnered widespread interest in hybrid and composite materials that include high-dielectric or charge-trapping inorganic components. In the polymer matrix incorporation of two dimensional (2D) material and inorganic nanoparticles can effectively (i) increase charge trapping and capturing capability, (ii) enhance the effective permittivity to boost stored charge, and (iii) alter surface roughness to enlarge real contact area^13,14. When appropriately balanced to prevent leakage, conductive polymers can simultaneously enhance charge transport to electrodes. Despite these alluring synergies, a significant obstacle still lies in creating ternary systems that take advantage of the complementing qualities of several fillers while preventing aggregation, phase separation, and mechanical stiffening¹⁵.

Polyaniline (PANI), titanium dioxide (TiO₂), and graphitic carbon nitride (g-C₃N₄) make up an attractive trio of useful elements for TENG electrodes. In exfoliated forms, g-C₃N₄ nanosheets have a high specific surface area, intrinsic trap states that can localize charge, and a large number of lone-pair electrons on nitrogen atoms. Because of these characteristics, when applied to the polymer surface, it functions as a surface modifier and a charge reservoir¹⁶. The chemical stability, polarization mediated by defects, high density of oxygen-vacancy-related trap states, and comparatively large dielectric constant of TiO₂ nanoparticles are well known; even very low loadings can raise the composite permittivity and produce deep traps that lessen charge recombination¹⁷.

Polyaniline (PANI) is a conducting polymer with strong electron-donating properties, which is due to abundant amino groups¹⁸. Protonation (doping) can be used to tune PANI, which provides polaron/bipolaron states that enable quick lateral conduction of charges to the electrode and enable charge extraction¹⁹. PANI offers regulated conductivity for charge collection, while g-C₃N₄ and TiO₂ provide trapping and dielectric enhancement when embedded in a compliant polymer binder such as PVDF. A potent tactic for promoting the enhanced TENG performance is to combine the special benefits of g-C₃N₄, TiO₂, and PANI to create effective tribo-positive materials. In the GTP ternary composite, protonated (amine-rich) PANI serves three complementary functions: (i) it creates a polaronic/conductive bridge between g-C₃N₄ and TiO₂, which promotes interfacial charge transfer and stabilises trapped charges (σ₀ t); (ii) it improves the composite’s local polarizability, favouring stored charge and Voc; and (iii) it is chemically and processing-compatible with PVDF (solvent, spin-coating, annealing), enabling a continuous, repeatable composite film²⁰. Regretfully, composite materials made of g-C₃N₄, TiO₂ and PANI have not yet been investigated for application in TENGs.

This systematic study addresses the tribo-positive layer in contact-separation TENGs by synthesizing and systematically characterizing a ternary g-C₃N₄/TiO₂@PANI (GTP) nanocomposite embedded in PVDF. In this work we have developed a facile and reproducible in-situ composite synthesis route that yields well-dispersed g-C₃N₄ nanosheets and TiO₂ nanoparticles attached with PANI matrix. After synthesis of ternary GTP nanocomposite was embedded in 3 wt% PVDF dispersion in dimethylformamide (DMF) by simple mixing method. g-C₃N₄, TiO₂, PANI, and ternary GTP nanocomposite’s structural, thermal and chemical properties (SEM, TEM, XRD, FTIR, DSC and UV-Vis) were analysed with triboelectric behaviour. And systematic investigation clarifies the ways in which each component—g-C₃N₄, TiO₂, and PANI—affects charge generation, trapping, and extraction in order to determine composition and microstructure windows that optimize electrical output and stability. We compare the device’s performance to that of PTFE-based tribonegative equivalents with different PTFE contents from 30 wt% PTFE to 60 wt% PTFE dispersion in Milli Q water. The primary novelty of this work lies in purposeful combination of three complementary functional fillers (g-C₃N₄, TiO₂ and PANI) within PVDF matrix and systematic characterization of structure, property and device relationships. Although single- or binary-component fillers have been demonstrated to improve TENG output in previous research, but there is limited thorough investigation of ternary nanocomposite systems, especially those that take advantage of both trapping (g-C₃N₄, TiO₂) and regulated conductivity (PANI) in a polymer binder. This work aims to provide mechanistic insight and useful guidelines for engineering composite tribo-layers for high-performance TENGs by combining advanced characterization with device-level electrical measurements (open-circuit voltage and short-circuit current). In the present study, the objective is not to benchmark absolute TENG output relative to pristine PVDF systems but rather to systematically investigate the influence of ternary g-C₃N₄/TiO₂@PANI interfacial engineering within the PVDF matrix and the effect of varying PTFE concentration in the tribonegative layer. Therefore, comparisons are restricted within the GTP–PVDF/PTFE compositional framework.

Experimental section

Materials

Aniline (C₆H₅NH₂), Hydrochloric acid (HCl), Ammonium per Sulphate (APS), Thiourea (CH₄N₂S), Dimethylformamide (DMF) {(CH₃)₂NC(O)H}, Isopropyl Alcohol (IPA) and Ethanol (C₂H₅OH) was purchased from LOBA Chemie, Mumbai INDIA. Polytetrafluoroethylene (PTFE) 60 wt% dispersion in Milli-Q water has been purchased from Sigma Aldrich. TiO₂ nanoparticles (rutile phase) purchased from Techinstro industries Maharastra, INDIA.

Synthesis of g-C₃N₄ nanosheets

The g-C₃N₄ nanosheets were produced using thiourea powder as a precursor. Five grams of thiourea were put in a round quartz crucible, enclosed with a lid, and heated for three hours at 500 °C in an air-filled muffle furnace. The yellow bulk g-C₃N₄ powder was gathered once the furnace had cooled. The bulk g-C3N4 was then heated for thermal exfoliation for three hours at 500°C in a muffle furnace to create g-C₃N₄ nanosheets.

Synthesis of ternary g-C₃N₄/TiO₂@PANI (GTP) nanocomposite and Polyaniline (PANI)

A straightforward in-situ chemical oxidative polymerization procedure was used to synthesize the ternary g-C₃N₄/TiO₂@PANI (GTP) nanocomposite, with the precursor weight percentage ratio being 70:25:5, respectively. After 30 minutes of ultrasonication, a homogenous mixture consisting of g-C₃N₄ (1g) and TiO₂ NPs (357mg) was prepared in a Milli Q water and ethanol (1:1) solvent. The dispersion was then moved to an ice bath (0–5 °C) with a 1 M HCl solution, and it was agitated for five minutes. Afterwards, 0.107g of aniline was added to the cold mixture. The polymerization process involved the dropwise addition of oxidant APS (315 mg in 15 ml water) while being continuously stirred. For three hours, the color changed from pale yellow to dark blue and was stirred. The solution was then filtered, extensively cleaned with water and IPA until it reached a pH between 6 to 7 (slightly acidic), and then dried at 70 °C in hot air oven. PANI was synthesized using a similar technique without the presence of TiO₂ and g-C₃N₄.

Preparation of g-C₃N₄/TiO₂@PANI (GTP) nanocomposite embedded PVDF dispersion

A homogenous suspension of PVDF embedded in GTP nanocomposite was made by dissolving 150 mg of PVDF powder in 5 mL of N, N-dimethylformamide (DMF) at 60 °C while being stirred magnetically until the powder was completely dissolved. After that, 100 mg of GTP nanocomposite was added to the PVDF solution, and stirred for 30 minutes, the mixture was constantly swirled to ensure that the nanoparticles were evenly distributed throughout the polymer matrix. In order to eliminate any agglomerates and provide a uniform nanocomposite solution appropriate for film preparation, the resultant suspension was further passed through a 0.45 μm PTFE syringe filter.

Preparation of 30wt%, 40wt%, 50wt% and 60wt% PTFE dispersions

Polytetrafluoroethylene (PTFE) dispersions were made in different weight percentages (30, 40, 50, and 60 wt%). The stock solution was a PTFE aqueous dispersion that was commercially available (60 wt% in water, Sigma-Aldrich). The stock dispersion was diluted using Milli-Q water (18.2 MΩ·cm) to reach the desired weight percentages, yielding the lower concentration dispersions (30 wt%, 40 wt%, and 50 wt%). Before use, every dispersion was well dispersed in order to ensure uniformity.

Spin-coating of triboelectric layers

Prior to spin coating, the ITO substrates were thoroughly cleaned by subsequent ultrasonication in detergent solution, Milli-Q water, isopropanol, and acetone to remove surface contaminants and improve film adhesion. The ITO-coated glass substrates were used as electrodes due to their smooth surface and high electrical conductivity, providing a rigid support for the active layer under low to moderate contact forces typical of low-power triboelectric energy harvesting. The deposited GTP–PVDF and PTFE films were thermally treated at 120 °C and 250 °C, respectively, to promote solvent removal and enhance interfacial bonding with the substrate. The resulting coatings exhibited stable adhesion during device assembly and repeated contact–separation operation without observable peeling or delamination. An NXG-P1 spin coater was used to fabricate all the functional films for the TENG device. The substrates were commercially available 2.5 × 2.5 cm² ITO coated glass slides (Sigma-Aldrich), with specification of surface resistivity 70-100 Ω/sq, and transmittance 86% (nominal at 550nm). In order to create electrical contacts, silver conductive ink was applied to the exposed ITO surface and allowed to cure in a hot air oven. To avoid coating, Kapton (polyimide) tape was used to mask the ITO substrate’s contact area. The NXG-P1 spin coater was then used to spin-coat the produced nanocomposite solutions onto the ITO substrates that had been cleaned. Cross-section SEM was utilised to determine the thickness of prepared tribo-positive and negative layers (films) coated on a clean glass substrate. The average thickness of the spin-coated GTP mixed PVDF composite positive layer is 11.7 ± 2.1 μm (n = 5 measurement points) and the PTFE negative layer is 0.7 μm (700 ± 40 nm, n = 5).

Fabrication of the electrodes for TENG device on ITO substrates

On cleaned ITO substrates, the tribopositive (GTP/PVDF) and tribonegative (PTFE) layers were applied using a two-step spin coating procedure with intermediate heat treatments. The most suitable film manufacturing parameters are compiled in Table 1.

Table 1.

Spin coating and heat treatment parameters for electrode fabrication.

Step	Volume (μL)	Speed (rpm)	Time (s)	Acceleration (s)	Thermal treatment
1	100	500	30	10	90 °C, 5 min (soft curing)
2	100	3000	60	10	120 °C for 15 min (PVDF), 250 °C for 15 min (PTFE)

The GTP nanocomposite-embedded PVDF suspension in DMF was deposited for the tribo-positive electrode using the aforementioned procedure, and it was finally cured for 15 minutes at 120 °C. Using the same spin coating parameters, aqueous PTFE dispersions at concentrations of 30 wt%, 40 wt%, 50 wt%, and 60 wt% were individually coated for tribo-negative electrodes. The films were then thermally annealed for 15 minutes at 250 °C to maximize surface charge density and film characteristics (Fig. 1).

Fig. 1.

Pictorial diagram of synthesis procedure of PANI and g-C₃N₄/TiO₂@PANI (GTP) nanocomposite.

TENG device assembly

In order to achieve controlled vertical displacement, the TENG device has been assembled together in a vertical contact-separation mode arrangement using two plastic platforms supported by four helical springs placed at the corners. The tribo-positive and tribo-negative electrodes were kept 3 cm apart to provide sufficient contact-separation movement while the device was in function. With a free length of 4 cm, an outer coil diameter of 5 mm, and a wire diameter of roughly 0.8 mm, the springs used offered reliable contact-separation behavior and structural stability even after repeated cycling. The plastic platforms were attached to the ITO substrates with deposited triboelectric coatings so that the active surfaces faced one another. Figure 2 shows the overall configuration of the TENG device as well as the electrode structure.

Fig. 2.

Design and structure of GTP nanocomposite embedded PVDF/PTFE based TENG device.

Total four TENG devices has been prepared in the series of 30 wt%, 40 wt%, 50wt% and 60 wt% of PTFE dispersion against ternary nanocomposite g-C₃N₄/TiO₂@PANI (GTP) embedded in PVDF (GTP-PVDF/PTFE based TENG). All triboelectric experiments were performed at a temperature of roughly 23 to 25 °C. Although active humidity control was not implemented, the relative humidity was kept within the normal range of ~45–55% for indoor summers.

Open-circuit voltage signals were analysed using a cycle-resolved approach. Individual contact–separation cycles were automatically segmented using a valley-to-valley detection algorithm to ensure complete capture of each electrical pulse. For each segmented cycle, the peak open-circuit voltage was defined as:

where represents the instantaneous open-circuit voltage. All reported voltage values in this work refer to unless otherwise specified. The short-circuit current signals were processed using the same cycle-resolved segmentation procedure. For each segmented cycle, the peak short-circuit current was defined as

Where I(t) represent the instantaneous current signal.

Statistical data analysis

Data processing and analysis

In order to preserve optimum measurement circumstances and suitable impedance matching for each signal type, electrical output signals, such as voltage and current, were measured in independent experimental runs. This method is consistent with established procedures for characterizing triboelectric nanogenerators²¹. Using pandas, numpy, and scipy tools, a custom Python analysis script was created to handle raw data with automatic cycle-level quantification, maintaining analytical consistency and reproducibility across datasets.

Signal pre-processing

Raw time-series data was processed systematically before analysis. In order to correct sampling abnormalities such as duplicate timestamps and non-monotonic time steps, time vectors were processed and standardized to a strictly monotonic seconds-based format. Such anomalies are known sources of artifacts in numerical differentiation and integration processes carried out on high-impedance signals, which is why these modifications were required²². To maintain signal integrity throughout the characterization procedure, separate analytical pipelines were used to process the voltage and current datasets.

Cycle detection and segmentation

An automatic valley-to-valley identification technique was used to segment individual electrical cycles. This approach ensures that every extracted cycle includes a full active pulse without truncation by locating local minima in the signal that match to boundaries between subsequent contact-separation events. In triboelectric nanogenerator waveform analysis, valley-to-valley segmentation techniques of this kind are frequently used, especially for measuring per-cycle electrical outputs under non-periodic or transient excitation conditions²³. An interquartile-range outlier detection criterion was used to rigorously reject non-physical cycles caused by electrical breakdown, irregular sparking, or acquisition artifacts in order to assure data quality. The interquartile range is defined as IQR = Q3 − Q1, where Q1 is the metric distribution’s first quartile (25th percentile) and Q3 is its third quartile (75th percentile). Cycles with metric values beyond the range [Q1 − 1.5×IQR, Q3 + 1.5×IQR] were identified as outliers and eliminated from further examination²⁴. Note that the transmitted electrical charge, which is represented separately in this work, should not be confused with the quartile notation Q1 and Q3, which only pertains to statistical percentiles used for outlier identification.

Metric calculation

For open-circuit voltage measurements, the peak voltage (Voc, peak) was directly calculated from each segmented voltage cycle. Short-circuit current and transferred charge measurements required more sophisticated processing to address baseline drift and spurious charge accumulation, which are well-documented challenges in triboelectric nanogenerator current characterization^25,26. Three successive phases were included in the current analysis methodology to ensure precise charge quantification. To eliminate systematic offsets, a cycle-specific baseline current was first calculated by calculating the median current value during each cycle’s quiescent pre-pulse region. This baseline was then removed from the entire waveform. Second, the temporal interval where the absolute current magnitude above five percent of the peak pulse amplitude was isolated to determine the active pulse window. By removing inactive areas where drift predominates, this threshold-based window definition assures that charge integration only records the actual triboelectric response. Third, the baseline-corrected current over the specified active window was numerically trapezoidally integrated to determine the transferred charge per cycle (ΔQ). In order to produce stable and repeatable cycle-resolved charge estimations, this pulse-resolved integration technique efficiently suppresses non-physical charge drift that builds up during inactive intervals between contact-separation events²⁷. Because cumulative drift can seriously skew charge quantification if ignored, the methodology is especially important for measurements carried out under varied contact circumstances or long sample times.

Statistical reporting and visualization

From time segments exhibiting dense and sustained cycle activity, representative waveforms were automatically picked. Following phase normalization, cycle-ensemble waveform statistics are displayed as median waveforms with interquartile range shading. In order to allow morphological comparison independent on tapping rate variations, current pulses underwent further peak alignment²⁸. Instead of reporting the mean and standard deviation, all distribution and stability graphs display the sample size, median, and interquartile range. Because manual tapping conditions naturally result in sporadic extreme events that would disproportionately affect parametric statistics, this option of robust statistics is suitable²⁹. Under these experimental conditions, the median and interquartile range offer more trustworthy measurements of central tendency and spread because they hold steady even in the presence of outlier events during lengthy measurement sequences. It should be noted that in order to preserve optimal measurement circumstances for each type of signal, voltage and current measurements were carried out in several experimental runs. As a result of the separate acquisition and processing paths used for each measurement modality, the number of approved cycles following quality filtering varies across voltage datasets and current or charge datasets.

Characterization techniques

The structural, morphological, optical, and electrical properties of the produced PANI and GTP nanocomposites were assessed using a thorough characterization process. A Bruker AXS D8 Focus diffractometer was used to obtain X-ray diffraction (XRD) patterns over a 2θ range of 10–80° using Cu Kα radiation (λ = 1.54 Å). Shimadzu UV1800ENG spectrophotometers were used to record UV–Vis absorption spectra in the 200–800 nm wavelength range. A PerkinElmer Spectrum 65 FTIR spectrophotometer was used to obtain Fourier transform infrared (FTIR) spectra between 4000 and 400 cm⁻¹. The thermal stability of the sample was determined through differential scanning calorimetry (DSC). Field emission scanning electron microscopy (FESEM), energy-dispersive X-ray spectroscopy (EDX), and elemental mapping with a JEOL, JSM-7900F microscope were used to examine the surface morphology and elemental composition. Using a JEOL JEM-2100 microscope, high-resolution transmission electron microscopy (HRTEM) pictures were captured. A Thermo Fisher Scientific device was used to measure X-ray photoelectron spectroscopy (XPS) in order to analyze the elemental and chemical states. A Keithley 2604B source meter was used for all electrical characterisation measurements.

Result and discussion

Synthesized g-C₃N₄, PANI and ternary nanocomposite g-C₃N₄/TiO₂@PANI (GTP) characterization

Optical properties analysis with UV-vis spectroscopy

UV-vis spectroscopy was used to analyze the optical properties of g-C₃N₄, TiO₂ NPs, PANI, and the ternary nanocomposite g-C₃N₄/TiO₂@PANI (GTP), as displayed in (Fig. 3). For optical band gap calculation, the most common and useful Tauc plot method was used for g-C₃N₄, TiO₂ NPs, PANI, and the ternary nanocomposite g-C₃N₄/TiO₂@PANI to study the absorption spectra. (Fig. 3a) displays the absorption spectra of pure g-C₃N₄ nanosheets, and the inset image shows the Tauc plot of g-C₃N₄ nanosheets for the direct bandgap energy. The optical band gap (E_g) has been calculated from following equation 1.

1

Fig. 3.

(a-d) UV-Vis absorption spectra of (a) g-C₃N₄, (b) TiO₂ NPs, (c) Polyaniline (PANI), (d) ternary nanocomposite g-C₃N₄/TiO₂@PANI (GTP). Inset displays Tauc plots for optical band gap (E_g). (e) Bar graph summarizes the band gap energy (E_g) values, (f) Urbach (disorder) energy analysis for GTP nanocomposite.

where α is absorption coefficient and α₀ is constant, hν is photon energy and E_g is optical band gap. According to Tauc’s plot, the change in the absorption edge represents the optical band gap. When (αhv)² is plotted against photon energy (hv), a straight line with an intercept equal to the material’s optical band gap is produced. The optical band gap is provided by the intercepts on the energy axis of the extrapolating Tauc curve in the linear section. Two unique optical characteristics of graphitic carbon nitride are visible in the UV-Vis absorption spectra of g-C₃N₄ nanosheets, which are shown in (Fig. 3a). The π→π* electronic transitions of the conjugated aromatic system inside the tri-s-triazine (heptazine) structural units are responsible for the absorption band that is 320 nm³⁰. Additionally, there is a noticeable absorption tail that reaches into the visible spectrum. This is the intrinsic band-edge absorption of g-C₃N₄, which results from n→π* transitions of the nitrogen atoms’ lone pairs. The Tauc plot in the inset of (Fig. 3a) was used to quantitatively measure the optical band gap. A band gap value of around 2.48 eV was obtained by linearly extrapolating (αhv)² vs photon energy³¹.

The produced TiO₂ nanoparticles’ UV-Vis absorption spectrum is shown in (Fig. 3b), where there is a strong absorption starting approximately 270 nm and high absorption in the UV region. By linearly extrapolating the plot of (αhv)² versus photon energy for a direct allowed transition to the energy axis, the optical band gap energy (E_g) was calculated from the Tauc plot displayed in the inset. This plot produced a band gap of approximately E_g 3.48 eV the observed value is well supported with the former TiO₂ nanoparticles³².

The UV-Vis absorption spectrum of polyaniline in the form of emeraldine salt is shown in (Fig. 3c). Two unique absorption features that are typical of doped polyaniline are visible in the spectrum. The π→π* electronic transitions of the benzenoid rings in the polymer backbone are responsible for the first sharp absorption peak, which is located at about 245 nm. This transition is typical of aromatic conjugated systems and correlates to the HOMO-LUMO gap of the benzenoid segments³³. And second is a noticeable broad absorption band in the visible spectrum that is centered at around 462 nm. It is caused by the polaron-to-π* band transitions and the benzenoid-to-quinoid excitonic transition. The green hue of conducting PANI is caused by this absorption, which is specific to the emeraldine salt form³⁴. The optical band gap was estimated to be around 2.20 eV based on the Tauc plot as displayed in the inset of (Fig. 3c). The prolonged π-conjugation along the polymer chain and the existence of low-energy polaronic transitions in the doped state are compatible with this comparatively narrow band gap³⁵.

As presented in (Fig. 3d), a broad visible band near 628 nm and a UV absorption peak around 332 nm are two unique optical characteristics of the GTP composite. The UV absorption at 332 nm is mainly caused by band-edge transitions in g-C₃N₄ and TiO₂, with TiO₂ contributing more in the UV region³⁶. The observed peak at about 628 nm corresponds to interfacial charge-transfer (CT) states that develop between PANI and the inorganic phases or between g-C₃N₄ and TiO₂, as well as to polaron and bipolaron absorptions in PANI³⁷. The observed optical band gap (E_g) of 3.18 eV (inset Fig. 3d) is more comparable to that of TiO₂ than g-C₃N₄ because the strong UV contribution from TiO₂ is primarily responsible for the high density of near-edge absorption³⁸. Bar graph in (Fig. 3e) summarizes the optical band gap (E_g) 2.48 eV, 3.48 eV, 2.20 eV and 3.18 eV for g-C₃N₄, TiO₂ NPs, PANI, and GTP nanocomposite respectively. A distinctive low-energy absorption tail that extends past the main band edge can be seen in the GTP ternary nanocomposite’s UV-Vis absorption spectra (Fig. 3d) Urbach-type exponential band-tail states are indicated by this tail, which continues into sub-bandgap photon energies³⁹. These band-tail states are caused by structural and electrical disorder, including as surface states, interfacial imperfections, lattice defects, and localized states at the composite’s heterogeneous interfaces⁴⁰. The degree of disorder and the breadth of the band-tail state distribution are measured by the Urbach energy (E_U). By using the empirical Urbach rule to analyse the exponential absorption edge, it was ascertained by following equation 2.

2

where α is the absorption coefficient, α₀ is a constant, hv is the photon energy, and E_g is a band gap energy. (Fig. 3f) displays the GTP nanocomposite’s ln(α) against photon energy (hv) curve in the sub-bandgap region. A linear fitting examination of the exponential tail area yielded a slope of 3.22921 eV⁻¹. An Urbach energy of 0.309 eV was calculated. This comparatively high Urbach energy suggests that the composite architecture has significantly increased structural and electrical disorder⁴¹. Polymer-inorganic interactions at the PANI interface, surface states on inorganic nanoparticles, and interfacial defects at the g-C₃N₄/TiO₂ heterojunction are the three main causes of the disorder⁴². The band gap may have a high density of localized band-tail states that can act as charge trapping sites, as indicated by the significant E_U value⁴³. Although these trap states can increase the amount of charge that can be stored, which is advantageous for triboelectric applications, they can additionally make it more straightforward for non-radiative recombination to occur if there are electronic transport channels, which could lower the device’s overall efficiency⁴⁴.

FTIR, XRD and DSC analysis

FTIR spectra of TiO₂, g-C₃N₄, PANI and the GTP nanocomposite samples were recorded; to investigate alterations in the local bonding environment, the complete and magnified between 400 to 2000 cm⁻¹ spectra are displayed in Fig. 4a and b, respectively. FTIR spectrum shows distinctive vibrational bands of TiO₂ nanoparticles in fingerprint region. Vibrational modes of the Ti–O bond in the TiO₂ crystal lattice are responsible for the broad absorption band in the 700–430 cm⁻¹ region. In particular, the rutile phase’s characteristic Ti–O stretching vibration can be seen at about 517 cm⁻¹. This band at 434 cm⁻¹ is associated with the deformative vibration of the Ti–O bending mode⁴⁵. Furthermore, the Ti–O–Ti bridging links, which are indicative of the TiO₂ network structure, are blamed for the strong and wide band seen at 700 cm⁻¹⁴⁶. Bare g-C₃N₄ nanosheets’ FTIR spectra show distinctive vibrational bands that validate the graphitic carbon nitride structure (Fig. 4b). Since triazine (or tri-s-triazine) units are the basic building blocks of g-C₃N₄, the absorption peak at 800 cm⁻¹ corresponds to their distinctive breathing mode⁴⁷. The aromatic heterocyclic rings’ C=N double bond stretching vibration is attributed to the peak at 1562 cm⁻¹, whilst the aromatic C–N stretching vibrations are associated with the peak at 1246 cm⁻¹⁴⁸. Polyaniline has strong characteristic peaks at 1562 cm⁻¹, 1465 cm⁻¹, 1246 cm⁻¹, and 1129 cm⁻¹ (Fig. 4b), respectively^49,50. The characteristic peaks of the PANI backbone are the N=Q=N (Q, quinonoid structure) and N–B–N (B, benzenoid structure) stretching modes, which corresponded to the peaks at 1562 cm⁻¹ and 1465 cm⁻¹, respectively; the peaks at 1246 cm⁻¹ are attributed to the C=N stretching vibration peaks of aromatic amines⁵¹. The observed shifts and broadening of the 1200–1600 cm⁻¹ bands in the GTP nanocomposite as compared to pristine PANI and g-C₃N₄ confirm the presence of hydrogen bonding, π–π interactions, and electronic coupling, all of that support in the emergence of interfacial states⁵². PANI polaronic bands and C–N/C=N vibrations are tightly correlated with the appearance or strengthening of peaks around 1246 cm⁻¹, 1314 cm⁻¹, 1129 cm⁻¹, 1465 cm⁻¹, and 1562 cm⁻¹. This indicates that PANI retains its doped conductive form and related polaronic properties within the composite. Furthermore, the g-C₃N₄ breathing mode’s drop in strength or smearing around 800 cm⁻¹ and the Ti–O bands around 500 cm⁻¹ imply improved particle dispersion, less local order, and increased surface engagement in the composite structure.

Fig. 4.

(a) FTIR spectra of TiO₂ NPs, g-C₃N₄, PANI, and ternary nanocomposite g-C₃N₄/TiO₂@PANI (GTP), (b) Enlarged FTIR spectra between 400 to 2000 cm⁻¹, (c) XRD pattern of PANI, g-C₃N₄, TiO₂ NPs, and GTP nanocomposite, (d) DSC graph of GTP nanocomposite.

To investigate the crystallographic information and phase identification of TiO₂ NPs, g-C₃N₄, PANI, and ternary nanocomposite g-C₃N₄/TiO₂@PANI (GTP) XRD characterization was employed, as shown in (Fig. 4c). The X-ray diffraction pattern of TiO₂ NPs has all the significant peaks in the 2θ range of 10°–80°, which precisely corresponds to the standard JCPDS card #75–1537 displayed in (Fig. 4c). The achieved major diffraction peaks at 27.14°, 35.82°, and 53.93° were attributed to the crystalline planes of (110), (101), and (211) respectively⁵³. The bare g-C₃N₄ exhibits the distinctive graphitic carbon nitride reflection in the (002) plane peak at about 27.14°⁵⁴. In accordance with polymeric morphology, PANI exhibits a broad, low-intensity halo that is amorphous or broad semicrystalline in nature. The GTP nanocomposite diffractogram shows noticeable peak broadening and decreased peak intensities particularly for g-C₃N₄ reflections. These features are frequently seen when the crystallite coherence length is reduced by intercalation and lattice strain, or when nanoparticles are widely distributed within or on top of a polymer matrix⁵⁵.

In the composite, g-C₃N₄ exhibits decreased structural ordering, as shown by widened and weaker diffraction peaks, but TiO₂ retains its crystalline rutile phase. PANI’s character is primarily amorphous. All of these findings are consistent with a structural model in which TiO₂ nanoparticles are distributed across a g-C₃N₄/PANI matrix, exhibiting partial structural disorder at the heterointerfaces and significant interfacial interactions.

In order to comprehend the material’s thermal stability and phase transitions, differential scanning calorimetry (DSC) was utilized to examine the GTP nanocomposite’s thermal behavior (Fig. 4d). When heated, the GTP nanocomposite undergoes two different endothermic transitions, according to thermal studies. Physically adsorbed water or residual solvent desorption is represented by the first endothermic peak seen at 83°C. On the other hand, low-temperature relaxation processes linked to local segmental motions within the polymer matrix might be the cause of this transformation⁵⁶. Polymer transitions are responsible for a second endothermic peak at 139°C, which may be related to glass transition processes or structural ordering and rearrangement in the composite⁵⁷. Depending on the kind of dopant, doped polyaniline usually experiences structural relaxations and de-doping processes in the temperature range of 100–200°C, therefore this peak could also be the result of dopant/counter-ion reorganization in PANI⁵⁸

FESEM, TEM, SAED, EDX and elemental mapping based morphology analysis

For surface morphological and compositional investigation FESEM, EDX and elemental mapping has been done for ternary nanocomposite g-C₃N₄/TiO₂@PANI presented in (Fig. 5). In Fig. 5 (a–d), field-emission scanning electron microscopy (FE-SEM) micrographs obtained at increasing magnifications from 10,000× to 35,000× (scale bars: 1 μm to 100 nm) demonstrate a rough and hierarchical surface morphology constructed from small spheroidal particles (TiO₂ nanoparticles) embedded on or attached to larger platelet-like domains, which is the g-C₃N₄/PANI matrix. This morphological structure is typical of nanocomposite systems in which layered organic-polymer matrices hold inorganic nanoparticles⁵⁹. The GTP nanocomposite’s RGB overlay map (C, N, O, Cl, Ti) for the combined elementals C, N, O, Cl, and Ti is displayed in (Fig. 5e). Ti (green) overlaps with C/N indicate the presence of nanoparticles in the organic matrix; overall co-localization indicates the dispersion of nanoparticles in the N-rich organic host. The RGB overlay map confirms that TiO₂ nanoparticles are distributed throughout the g-C₃N₄/PANI host by displaying Ti signals co-localized with the C/N matrix. EDS elemental mapping provides detailed insight into the elemental distribution within the GTP composite. The C map (red) in (Fig. 5f) illustrates how carbon is evenly distributed throughout the imaged region, indicating a continuous organic phase. Polymer-based matrices in nanocomposite systems have this uniform dispersion. The N map (cyan) in (Fig. 5g) indicates that N and C are co-localized, indicating the existence of a g-C₃N₄/PANI matrix. The close mixing of these nitrogen-rich organic components is confirmed by the overlapping distribution of C and N. The existence of TiO₂ is confirmed by the oxygen map (yellow) in (Fig. 5h), which strongly coincides with the titanium distribution. The various oxygen species found in XPS O 1s spectra are consistent with the measured oxygen coming from both the TiO₂ crystal lattice and surface hydroxyl/C–O functional groups inside the organic phase. The chlorine map (orange) in (Fig. 5i) shows a very low chlorine level. This very low Cl presence is attributed to thorough washing procedures during PANI synthesis, which successfully eliminated extra dopant ions. Strongly bonded counter-ions within the PANI structure may be associated with residual chlorine. The EDX spectrum of mapped area illustrated in (Fig. 5k), with the inset table presenting the atomic % and weight % of the constituent elements. C 40.9 wt%, N 25.0 wt%, O 20.1 wt%, and Ti 13.9 wt% are the elemental compositions obtained by EDX quantification over the mapped region. Chlorine is found close to the detection limit at about 0.1 weight percent, indicating that it is present in the composite in extremely low quantities. The major organic matrix, which consists of g-C₃N₄ and PANI, coincides with the high carbon and nitrogen content. The significant nitrogen weight percentage (25.0 wt%) is consistent with the anticipated N/C ratio in graphitic carbon nitride structures and conducting polymers doped with nitrogen. Both the TiO₂ nanoparticles and the oxygen-containing functional groups (hydroxyl, carbonyl) on the organic components contribute to the oxygen content (20.1 weight percent). The presence of TiO₂ nanoparticles in the composite framework is confirmed by the titanium content (13.9 weight percent). While trace residual chlorine may exist as firmly bound counter-ions within the PANI backbone, the nearly insignificant chlorine content (~0.1 wt%) suggests that dopant anions were effectively removed during the synthesis washing phases.

Fig. 5.

(a–d) FE-SEM micrographs of GTP nanocomposite at increasing magnification (e) RGB overlay elemental map (C = red, N = cyan, O = yellow, Cl = orange, Ti = green). (f–j) Individual EDX elemental maps of C, N, O, Cl, and Ti (k) Representative EDX spectrum and quantification (wt% / at%) for the mapped area.

According to transmission electron microscopy (TEM) analysis, the GTP nanocomposite exhibits a hierarchical nanoarchitecture, with nanocrystalline TiO₂ NPs scattered throughout and decorating thin g-C₃N₄ nanosheets integrated in a polyaniline matrix, (Fig. 6). (Fig. 6b, c) illustrates irregular, connective PANI regions that form continuous polymeric networks alongside sheet-like, low-contrast domains corresponding to g-C₃N₄ nanosheets in low- and medium-magnification TEM micrographs. Intimate interfacial contact between the inorganic and organic phases is demonstrated by the homogeneous distribution of higher-contrast, nearly spherical TiO₂ nanoparticles on these organic substrates. The selected electron diffraction pattern (SAED) of the GTP nanocomposite can be observed in (Fig. 6d). TiO₂ nanoparticles could be the cause of the concentric rings that are typical of polycrystalline or nanocrystalline structure in the SAED pattern of the GTP nanocomposite. Furthermore, amorphous or partially disordered organic components (g-C₃N₄ and PANI) are responsible for the diffuse halo that is observed. In hybrid organic-inorganic nanocomposites, where crystalline inorganic nanoparticles coexist with semicrystalline or amorphous polymer matrix, this combination of diffuse dispersion and crisp diffraction rings is typical.

Fig. 6.

TEM Images of the GTP nanocomposite. (a) TEM showing sheet-like domains interspersed with higher-contrast particles. (b) At 500 nm scale image identifying g-C₃N₄ sheets, PANI chains and TiO₂ nanoparticles. (c) At 100 nm scale image show interconnected PANI/g-C₃N₄ networks embedding TiO₂ crystallites. (d) SAED pattern of ternary nanocomposite GTP.

X-ray photoelectron spectroscopy (XPS) analysis of GTP nanocomposite

The chemical composition and interfacial bonding of the components present in the as-prepared GTP nanocomposite have been examined using X-ray photoelectron spectroscopy (XPS), as shown in (Fig. 7). The survey spectra of ternary nanocomposite g-C₃N₄/TiO₂@PANI (GTP) displayed in (Fig. 7b). Three separate peaks centred at 284.7, 286.6, and 288.6 eV are observed in the deconvoluted C 1s XPS spectra of the GTP nanocomposite (Fig. 7b), which reveals information about the carbon atoms’ chemical bonding conditions within the ternary nanocomposite. The predominant signal at 284.7 eV is associated with sp²-hybridized C–C/C=C bonds from the graphitic domains of g-C₃N₄ and the aromatic rings of polyaniline⁶⁰. Carbon atoms bound to heteroatoms (C–N⁺/C–O/C=N) within the protonated PANI backbone and at the g-C₃N₄–TiO₂ interface produce the intermediate peak at 286.6 eV, which may represent partial oxidation and electronic interaction between the organic and inorganic phases. The N–C=N species in the tri-s-triazine (heptazine) framework of g-C₃N₄ are responsible for the high-binding-energy peak at 288.6 eV, indicating that the structural integrity of graphitic carbon nitride is maintained in the composite⁶¹. Two separate components are visible at 398.7 eV and 400.1 eV in the deconvoluted N 1s XPS spectra (Fig. 7c). The strong signal at 398.7 eV is associated with sp²-hybridized nitrogen atoms (C–N=C) in the aromatic tri-s-triazine (heptazine) units of g-C₃N₄, indicating the preservation of the conjugated heterocyclic framework during composite synthesis⁶². Tertiary or protonated nitrogen species (C–N–H/C₃–N⁺H), which originate from doped polyaniline and edge or defect sites of g-C₃N₄, are responsible for the component at 400.1 eV⁶³. This peak may also comprise nitrogen coupled to titanium (Ti–N–C) at the organic–inorganic interface. These two nitrogen states show a hybrid bonding environment where g-C₃N₄ and PANI are chemically linked with TiO₂, allowing for enhanced electronic coupling and interfacial charge transfer pathways that could increase triboelectric performance.

Fig. 7.

XPS of ternary nanocomposite g-C₃N₄/TiO₂@PANI (GTP) (a) Survey spectra of GTP, High resolution XPS core-level spectrum graphs of (b) C 1s, (c) N 1s, (d) Ti 2p, (e) O 1s, (f) Cl 2p.

(Fig. 7d) demonstrates the high-resolution Ti 2p spectrum, which presents Ti 2p_3/2 at 458.7 eV and Ti 2p_1/2 at 464.4 eV (Δ = 5.7 eV). TiO₂ include Ti⁺⁴, which is compatible with the symmetric peak position of Ti 2p_3/2 at 458.7 eV⁶⁴. Three primary components were identified from the deconvoluted O 1s spectra of the GTP composite (Fig. 7e). Lattice O²⁻ in Ti–O–Ti (TiO₂) is responsible for the main peak at 529.7 eV. Surface hydroxyls are represented by the larger binding energy component at 532.3 eV, which may corrospond to Ti–O–C/C–O species as a result of interfacial interaction with the organic phases (PANI/g-C₃N₄)⁶⁵. Significant surface modification of TiO₂ by the polymer and carbon-nitride is indicated by the relative intensity of the defect and hydroxyl components. This is consistent with C 1s shifts (mid-peak at 286.6 eV) and the broad Urbach tail observed in optical spectra (Fig. 3f), both of which support the existence of charge-transfer states and increased trap density in the composite. As illustrated in (Fig. 7f), the high-resolution Cl 2p spectra was deconvoluted into three peaks with spin–orbit doublets (ΔE = 1.6 eV). In accordance with Cl⁻ counter-ions from HCl-doped polyaniline, the dominating 197 eV is attributed to ionic chloride (Cl⁻). A portion of more strongly bound chlorine (covalent C–Cl or Cl coordinated at the organic–inorganic interface) is indicated by a secondary component with a higher binding energy 198.4 and 200 eV⁶⁶. The maximum peak value attained in a representative performance cycle under repeated contact-separation excitation is consistent with the measured open-circuit voltage and short-circuit current.

Fabricated TENG device’s electrical characterization

The electrical output performance of the fabricated GTP-PVDF/ 30 wt% PTFE TENG, GTP-PVDF/ 40 wt% PTFE TENG, GTP-PVDF/ 50 wt% PTFE TENG, and GTP-PVDF/ 60 wt% PTFE TENG devices was systematically investigated. A total of four TENG devices have been fabricated. (Fig. 8a) shows continuous open-circuit voltage (Voc(t)) measurements obtained during subsequent device tests; each colored section denotes a different device composition. (Fig. 8b) shows the Voc(t) graph of the GTP-PVDF/60wt% PTFE-based TENG device, which has the highest peak voltage of 4.56 V among all TENGs (30wt% PTFE, 40wt% PTFE, and 50wt% PTFE). (Fig. 8c) displays the single contact separation cycle for releasing and pressing, and the response time of a 60 wt% PTFE TENG device. The fastest response time is 67 milliseconds (ms) for GTP-PVDF/60 wt% PTFE (Fig. 8e), while all variants, including 30 wt%, 40wt%, and 50 wt% PTFE-based GTP-PVDF/PTFE TENGs, have substantially slower responses in Voc. These TENG devices have response times of 72 ms, 87 ms, and 70 ms for 30 wt% PTFE TENG, 40 wt% PTFE TENG, and 50 wt% PTFE TENG, respectively. The faster response time at 60 wt% (67 ms) suggests faster charge separation may be due to contact mechanics at 60 wt% (and potentially less trapped mobile counter-ion motion that would impede the Voc rise). As demonstrated in (Fig. 8d), the maximum peak open circuit voltages of GTP-PVDF/PTFE-based TENGs (30 weight percent PTFE, 40 weight percent PTFE, 50 weight percent PTFE, and 60 weight percent PTFE) are 3.61V, 3.5V, 3.4V, and 4.56V, respectively. The open-circuit voltage (Voc) shows a composition-dependent variation with PTFE concentration, reaching a maximum value of 4.56 V at 60 wt% PTFE. This behavior suggests that increasing PTFE content does not necessarily lead to a proportional increase in Voc. Instead, the tribonegative layer likely reaches an optimal balance of surface morphology, triboelectric negativity, and reduced charge leakage at 60 wt%, which enhances the effective surface charge density and consequently the generated voltage. This observation can be understood within the framework of the TENG capacitor model and the intrinsic triboelectric properties of PTFE. Because PTFE possesses strong electronegativity in the triboelectric series, higher PTFE fractions promote increased negative surface charge generation during contact–separation. In addition, PTFE’s relatively low dielectric constant (εr ≈ 2.1) reduces the effective device capacitance according to the capacitor relation , where is the voltage, is the transferred charge, and is the capacitance. As indicated by Equations 3 and 4, the enhanced surface charge density (σ) associated with PTFE’s electron affinity increases the electric potential difference across the device.

3

4

Fig. 8.

(a) Open circuit voltage (Voc(t)) of GTP-PVDF/PTFE-based TENGs (in variation of PTFE from 30wt% to 60wt%), (b) Voc(t) of GTP-PVDF/60 wt% PTFE TENG device, (c) Maximum peak voltage of all TENG devices, (d) Response time for GTP-PVDF/PTFE (60wt%) based TENG for Voc(t), (e) Response time bar graph for all TENG devices for Voc(t).

Where σ is surface charge density, ε₀ is vacuum permittivity, ε_r is relative permittivity, d is the air gap at maximum separation, and A is the electrode area.

(Fig. 9a) illustrates continuous short-circuit current (Isc(t)) measurements obtained during subsequent device tests; each coloured section corresponds to a different device composition. (Fig. 9b) demonstrates a short circuit current (Isc(t)) graph of the 30 wt% PTFE base GTP-PVDF/PTFE TENG device, which has the highest peak current of 2.86μA among all TENGs (30 wt% PTFE, 40 wt% PTFE, 50 wt% PTFE, and 60 wt% PTFE). (Fig. 9c) presents the single contact separation cycle for releasing and peak response time of 30 wt% PTFE TENG device in short-circuit current Isc(t). A 30 wt% PTFE-based GTP-PVDF/PTFE TENG device has a response time of 26 ms despite having a lower Isc(t). Fastest response time in short circuit current is 17 ms for 60 wt% PTFE-based TENG. Peak response time can be seen in the bar graph (Fig. 9e) for all TENG devices. GTP-PVDF/PTFE TENGs (30 wt% PTFE, 40 wt% PTFE, 50 wt% PTFE, and 60 wt% PTFE) have maximum peak short circuit current as a function of time of 2.86μA, 1.20 μA, 1 μA, and 0.91 μA, respectively, as shown in (Fig. 9d). Bar graph in (Fig. 9f) demonstrates the peak to power (P = VI) for 30 wt% PTFE, 40 wt% PTFE, 50 wt% PTFE, and 60 wt% PTFE TENGs. The highest peak power is 10.32μW for a 30 wt% PTFE TENG device.

Fig. 9.

(a) Short-circuit current output (Isc(t)) of GTP-PVDF/PTFE based TENGs variants (30 wt%, 40wt%, 50wt% and 60wt%), (b) Single spectra of short-circuit current Isc(t) for 30 wt% PTFE based TENG device, (c) Response time for 30 wt% PTFE TENG device, (d) Maximum peak current (Isc(t)) of all TENG devices, (e) Response time bar graph for all TENG devices for Isc(t), (f) Maximum peak power in bar graph for all TENG devices.

The maximum Voc is produced by the 60 wt% PTFE-based TENG, whereas the 30 wt% PTFE TENG produces the maximum Isc and peak power. This phenomenon can be understood in terms of PTFE’s extreme negative triboelectric polarity, its low dielectric constant, and the composite’s interfacial properties. When the electrodes come into contact or separate, the 30% PTFE based TENG with PVDF and conductive GTP (g‑C₃N₄/TiO₂@PANI) nanocomposite provides channels that may exchange charges more quickly, resulting in high Isc⁶⁷. Specifically, PTFE easily collects surface charge upon contact and has a very high electronegativity. Therefore, films with higher PTFE amount hold more net surface charge, boosting Voc, as explained by equations 3 and 4. Conversely, compared to the highly PTFE-rich films, the 30% PTFE composite has a greater effective permittivity and better charge mobility due to the fact that it contains more PVDF (ε ≈8–10) and GTP filler (g–C₃N₄/TiO₂@PANI)⁶⁸. The GTP nanocomposite generates several deep interfaces where charges can localize. Charge recombination is slowed by these interfacial states, also known as trap states. In fact, current research demonstrates that deep charge-trap sites (such as MOFs) in a TENG layer significantly increase retained surface charge and decrease charge dissipation⁶⁹. The doped PANI and g-C3N4/TiO2 semiconductor heterojunctions in the composite probably introduce trap states that trap ions or tribo-electrons. Since the composite’s trapping and insulating properties are at their peak at 60% PTFE, more charge is retained (increasing Voc), but charge movement is sluggish. Rapid charge transfer (increased Isc) before recombination is made possible by trap-mediated retention, which is still present but less prominent at 30% PTFE. The visualization is further supported by spectroscopic evidence; the GTP composite’s chemical bonding environment is confirmed by XPS spectra (C1s, O1s, N1s, Ti2p, and Cl2p). Subgap trap states in the GTP nanocomposite are consistent with a moderate bandgap and an extended Urbach tail in UV-Vis absorption’s Tauc-plot analysis of the GTP composite. More localized defect/trap states that are capable of capturing charges under friction are implied by a greater Urbach energy (tail). The 30 wt% PTFE TENG produces the highest current and power because higher permittivity generally boosts TENG output by increasing capacitance and charge production. Table 2 compares GTP–PVDF/PTFE TENG devices with their performance, architecture, and triboelectric materials with representative g-C₃N₄-based binary and ternary composites reported in the recent literature. The table highlights the distinction of the present ternary interfacial engineering approach relative to prior binary or conductive hybrid systems.

Table 2.

Output performance comparison with other TENG devices.

Sr. No.	Ref. (year)	TENG device	Composite/ Architecture	TENG mode	Performance	Key role of materials
1.	This work	GTP-PVDF/PTFE TENG	g-C₃N₄/TiO₂@ PANI (GTP) in PVDF + PTFE	Vertical contact separation (CS )	Voc = 4.56 V, Isc= 2.86 μA, ΔQ= 200.4 nC, Pmax=10.3μW	Unique ternary filler engineered for charge traps & charge transfer(CT) states and systematic PTFE variation
2.	Nishat et al. 202470	SILAR g-C₃N₄/CdS TENG	SILAR g-C₃N₄/CdS TENG	Contact Separation	Voc = 29.6 V; Isc = 0.55 μA; Power density =2.16 μW/cm2	Hybrid with light enhancement; CdS–g-C₃N₄ hybrid enhances CT, traps
3.	Y. Xiao et al. 202471	PVA/g-C₃N₄ hydrogel TENG	2D g-C₃N₄ in PVA hydrogel	Single electrode mode	Voc = 80 V, Isc = 2.5 μA, Power density =45 μW/cm2	g-C₃N₄ increases dielectric polarity & flexible contact
4.	Z. Yang et. al. 202472	g-C₃N₄/PVDF TENG	g-C₃N₄ in PVDF vs nylon-11	CS mode	Voc = 339 V; Isc = 14.5 μA; Power density = 0.94 W/m2	g-C₃N₄ enhances electronegativity & piezo-effect in PVDF
5.	V. S. Niranjana et al. 202573	Electrospun PVDF/Si-HBP-G1 TENG	Electrospun PVDF/Si-HBP-G1 vs Al	CS mode	Voc = 76 V, Isc = 2.1 μA, Power density = 0.035 W/m2	Si-HBP-G1 in PVDF increases trapping sites and increases charge density
6.	Dheeraj Kumar et. al. 202574	g-C3N4@BaTiO3 (CN@BT) P(VDF-TrFE) TENG	g-C3N4@BaTiO3 (CN@BT) P(VDF-TrFE) vs PET ITO	CS mode	Current density = 2.59 μA/cm2 , Power density = 0.56 mW/cm2	Photoactive ferroelectric ternary composite g-C3N4@BaTiO3 (CN@BT) P(VDF-TrFE) film-based TENG (Photo-FE@TENG)
7.	Y. Xiao et al. 202575	Ag/g-C₃N₄ & CNT/g-C₃N₄ PDMS TENGs	Ag–g-C3N4 or CNT–g-C3N4 in PDMS vs CNF	CS mode	Voc = 92-122V, Isc = 4.6-5.8 μA, Charge transfer (Qsc) =49-105 nC, Power =1.45 W/m2	Hybrid dopants optimize conductivity, traps, and contact. Hybrid designs focus on conductive 1D fillers.
8.	Ying Su et al. 202176	POM-modified g-C₃N4 TENG	POM@g-C₃N₄ on ITO/PET	CS mode	Voc = 78 V; Isc = 0.657 μA; Qsc = 15 nC	Binary trap engineering, POM increases work function & electron storage
9.	S. P. Das et. al. 202377	Au-g-C₃N₄/ZnO TENG	Au–g-C₃N₄/ZnO hierarchical on paper	Flexible CS TENG	Power = 3.5 μW/cm2	Hierarchical dielectric + surface roughness
10.	Maosen Yang et. al. 202378	PANI/cellulosic hierarchical structure (PC-TENG)	PANI/nitro-cellulose paper and nitro-cellulosic paper	CS mode	Voc = 23 V, Isc = 0.17 μA	self-powered gas sensors, detection of NH3

Crest factor calculation

To better understand the charge release behaviour of the GTP–PVDF/PTFE system, the current crest factor (CF) was evaluated on a cycle-resolved basis. The crest factor is defined as the ratio of the peak current to the root-mean-square (RMS) current within each cycle and provides a useful descriptor of waveform impulsiveness. The crest factor was calculated as⁷⁹:

where represents the maximum absolute current within the cycle and is the root-mean-square current over the same interval. Individual cycles were identified using valley-to-valley segmentation of the current signal, and outlier cycles were removed using a 1.5× IQR statistical filter before summarizing the results.

The device containing 30 wt% PTFE exhibited a median crest factor of 3.2 (IQR: 2.7–4.3, n = 113 cycles), whereas the 60 wt% PTFE device showed a higher median crest factor of 4.3 (IQR: 3.7–6.1, n = 34 cycles). The larger CF observed for the 60 wt% PTFE device indicates a more impulsive current waveform with sharper discharge peaks relative to its RMS level. In contrast, the 30 wt% PTFE device exhibits a lower crest factor and a smaller interquartile range, indicating more distributed current pulses and improved cycle-to-cycle stability. This behavior is consistent with the higher short-circuit current and power output observed for this composition. Because triboelectric nanogenerators generate electrical output through displacement current induced by periodic electrostatic potential variation during the contact–separation process, their electrical signals typically appear as transient current pulses rather than steady conduction currents³. Furthermore, material-dependent variations in triboelectric charge density can influence electrostatic field development and discharge behavior, as demonstrated in quantitative studies of the triboelectric series⁸. These observations, therefore, provide waveform-level evidence supporting the voltage–current trade-off governed by PTFE composition within the GTP–PVDF/PTFE triboelectric system.

Statistical analysis of 30 wt% and 60 wt% PTFE-based TENGs ’ electrical data for stability and repeatability

Figures 10, 11, and 12 are arranged to progress from representative raw waveforms to ensemble statistics and finally to cycle-to-cycle stability, enabling a clear, non-redundant assessment of the effects of PTFE composition.

Fig. 10.

Representative output waveforms of TENG devices with different PTFE weight fractions. (a–c) Open-circuit voltage , short-circuit current , and transferred charge obtained for the 60 wt% PTFE device under manual tapping. (d–f) Corresponding representative voltage, current, and charge waveforms for the 30 wt% PTFE device recorded under comparable experimental conditions. Voltage and current signals were acquired in separate experimental runs optimized for each measurement and are therefore not time synchronized. The transferred charge was calculated by numerical integration of the baseline-corrected current pulse over the active conduction window, and denotes the net per-cycle charge transfer for the representative cycle shown.

Fig. 11.

Cycle-ensemble waveform statistics for different PTFE compositions. (a) Ensemble statistics of voltage and current cycles for the 60 wt% PTFE device, plotted on a normalized phase axis (0–1). Solid lines represent the median cycle, and shaded regions denote the interquartile range (IQR) across all quality-controlled cycles. (b) Ensemble statistics of voltage and current pulses for the 30 wt% TENG device. Current cycles were peak-centered (peak aligned at phase = 0.5) prior to phase normalization to enable comparison of pulse morphology independent of tapping rate. Median and IQR bands summarize the distribution of pulse shapes for each composition.

Fig. 12.

Cycle-to-cycle stability and statistical distributions of output metrics. The identical analysis and visualization protocol was applied to both PTFE compositions—60 wt.% (a) and 30 wt.% (b) enabling a direct and unbiased comparison of electrical stability and performance distributions. Cycle-resolved stability and violin plots illustrating the variation of peak voltage, peak current, and per-cycle transferred charge across consecutive contact–separation cycles. Each data point corresponds to one automatically detected electrical cycle, highlighting the repeatability and temporal stability of the device output under manual tapping conditions. In the violin plots, the central line denotes the median value, while the width represents the probability density of the measured data. Non-physical cycles arising from sporadic sparking or acquisition artifacts were excluded using a 1.5× interquartile range (IQR) criterion. Reported sample sizes correspond to post-quality-control datasets used for statistical analysis.

Individual contact-separation events transfer tens of nanocoulombs of charge every pulse, as shown by representative time-domain measurements. For a representative occurrence (Fig. 10 c), numerical integration of the current trace of 30 wt% PTFE TENG results in a transferred charge of ΔQ = 25.0 nC. The related voltage and current waveforms for 30 wt% PTFE TENG (Fig. 10a,b) show that peak Isc(t) appear as short microsecond-scale spikes with microampere-level peak amplitudes that coincide with mechanical contact, whereas peak Voc(t) reaches roughly 3 V. The analytical contact-separation mode triboelectric nanogenerator model is in line with these findings. The immediate electrode gap distance and the net stored charge control the open-circuit voltage, whereas the charge transfer rate (dQ/dt) and contact mechanics control the electrode separation’s temporal evolution and control the short-circuit current. While the current pulses record the quick charge redistribution that takes place during contact and separation transitions, the voltage waveform depicts quasi-static charge storage dynamics.

For representative contact-separation occurrences, the integrated short-circuit current results in a transferred charge of ΔQ = 200.4 nC per pulse (Fig. 10f). Peak open-circuit voltages around ≈3.3 V and μA-scale current spikes (I = 1.0–1.2 μA) are displayed in simultaneous voltage and current windows (Fig. 10d, e). The integrated charge plateau, which is consistent with trap/charge-transfer states in the GTP filler that increase retained charge, shows that the accessible transferable charge was saturated during that event.

A clear materials-dependent trade-off is revealed by ensemble cycle analysis. The voltage envelope for the 60 wt% PTFE device (Fig. 11a) exhibits a wide IQR and a persistent median voltage through separation, whereas the current pulses are tiny but steady across time. The 30 wt% PTFE device (Fig. 11b), Contrary, generates significantly larger median current pulses and increased pulse-to-pulse amplitude, indicating more efficient transient charge transfer. These patterns can be explained by competing effects: smaller PTFE fractions result in softer contacts and higher effective capacitance, which boost dQ/dt and short-circuit current, while increasing PTFE raises triboelectric surface charge (favouring higher retained Voc).

A distinct composition-dependent trade-off between voltage amplitude and charge transfer capacity in electrical output characteristics is shown by the cycle-resolved study. The 60 wt% PTFE device often produces higher peak voltages, with median values in the 3.6–3.8 V range and sporadic excursions to roughly 4.5 V under vigorous contact conditions, as seen in Figures 10, 11, and 12. The 30 wt% PTFE device, on the other hand, generates significantly more current and transmitted charge per cycle but has slightly lower median voltage levels, usually 3.2–3.4 V. This redistribution implies that the triboelectric interface is altered by lower PTFE loading in a way that prioritizes charge mobility and transfer efficiency over voltage accumulation. As the composite microstructure changes with filler content, variations in surface charge trapping density, interfacial capacitance, or charge dissipation pathways may be reflected in the measured voltage-charge trade-off. Optimization for certain application requirements is made possible by this composition-tuneable output distribution: optimum charge transfer for energy harvesting or actuation against high-voltage production for sensing applications.

This increase in charge transfer is consistently demonstrated by several types of analytical methods. Time-domain studies reveal that the 30 wt% PTFE device generates higher peak current amplitudes, with median values of around 1.1–1.3 µA compared to 0.3–0.5 µA for the 60 wt% PTFE device. The transmitted charge per cycle shows an even more striking compositional dependence, with median values of over 200 nC for the 30 wt% PTFE device—nearly an order of magnitude higher than the 20–30 nC measured for the 60 wt% PTFE composition. Complementary visualization techniques validate these tendencies’ resilience. Consistent composition-dependent behavior can be shown in representative waveforms (Fig. 10), ensemble median profiles with interquartile range shading (Fig. 11), and stability plots with violin distributions (Fig. 12). This multifaceted agreement indicates persistent changes in triboelectric performance, rather than occasional high-amplitude outlier occurrences, are responsible for the observed increase in current and charge output for the 30 wt% PTFE device. The conclusion that PTFE loading systematically modifies the voltage-charge trade-off in these composite triboelectric surfaces is strengthened by the agreement between various analytical viewpoints. This study’s observation of the decoupling between voltage amplitude and charge transfer is in line with the behavior of triboelectric nanogenerators, which has been extensively documented to show different responses to changes in material composition, device geometry, and operating conditions^3,21. The current findings show that while lowering PTFE loading improves charge-dominant performance, increasing PTFE content favors voltage-dominant output, as shown by higher integrated charge per cycle and current amplitudes in the 30 wt% PTFE device.

Conclusion

We report a successful synthesis of ternary nanocomposite GTP (g-C₃N₄/TiO₂/PANI) and GTP-enabled PVDF/PTFE TENG whose performance is governed by a balance between surface charge generation and dynamic charge transfer. Structural and spectroscopic analyses (TEM/SEM, XPS, UV–Vis, EDX) confirm well-dispersed TiO₂ nanoparticles anchored on g-C₃N₄ sheets and bridged by polyaniline, producing a high-surface-area, N-rich organic matrix with abundant interfacial sites. Protonated N and PANI-related species (C 1s at 286.6 eV, N 1s at 400.1 eV), Ti–O–C/Ti–N interfacial bonding signals, and a small Ti³⁺/oxygen vacancy population are all preserved, according to XPS investigation. Polaronic and sub-bandgap absorption characteristics with a high Urbach energy (E_U = 0.31 eV) are revealed by UV-Vis spectroscopy, suggesting significant trap states and charge transfer pathways. Polarizability and charge retention at the organic–inorganic heterojunction interfaces are improved by these features. A basic trade-off between Voc and Isc output is shown by electrical measurement of TENGs with different PTFE contents (30–60 wt%). Due to PTFE’s high surface charge density and low dielectric constant, which boost surface potential, the 60 wt% PTFE device reaches the maximum open-circuit voltage (Voc = 4.56 V). On the other hand, due to better mechanical contact, increased effective permittivity, and faster charge transfer kinetics, the 30 wt% PTFE device generates the maximum short-circuit current (Isc(t) = 2.86 μA) and peak power under matched load conditions (Pmax = 10.32 μW). Best performed TENG device’s (30wt% and 60wt% based TENG) were further analysed by statistical calculations. In order to create interfacial electronic states and surface morphology for improved triboelectric nanogenerator performance, this work presents a scalable materials design technique that combines g-C₃N₄, TiO₂, and doped PANI.

Author contributions

Ashok Kumar Swami : Manuscript Writing, conceptualized and designed the study, coordinated the research activities, and supervised data interpretation, Nelapati Avinash: characterisation, Deepak Verma : reviewed and edited the manuscript, conceptualized and designed the study

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

All data generated or analysed during this study are included within the manuscript. No additional data or supplementary files are available.

Declarations

Competing Interests

The authors declare no competing interests.

Declarations of Generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the author(s) used AI tools to improve the readability of the manuscript/ language refinement. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the published article.