A high output triboelectric nanogenerator based on 2D boron nitride nanosheet–PVP composite ink and electrospun cellulose acetate nanofibers for kinetic energy harvesting and self-powered tactile sensing applications

Abstract

The development of intelligent systems integrated with high-sensitivity sensors is critical for next-generation electronic applications. Triboelectric nanogenerator (TENG)-based tactile sensors offer a promising solution by converting mechanical stimuli directly into electrical signals, making them ideal for wearable electronics, robotics, and prosthetics. In this work, we present a self-powered tactile sensor fabricated using two complementary triboelectric materials: screen-printed boron nitride nanosheet (BNNS) composite ink printed on a polymer substrate and electrospun cellulose acetate (ES-CA) nanofibers. Structural modification of the BN–PVP/ES-CA TENG resulted in a significantly enhanced performance, delivering an output voltage of 1200 V, a short-circuit current density of 1.2 mA m⁻², and a power density of 1.4 W m⁻². The sensor effectively detects low-magnitude forces even up to 0.05 N, exhibiting a sensitivity of 3.98 V N⁻¹ for forces <2 N and 1.843 V N⁻¹ for forces between 2 and 10 N, demonstrating its potential in high-resolution tactile sensing for advanced robotic and prosthetic applications.

---

1. Introduction

Low-cost sensor arrays form the backbone of modern day technologies such as Internet of Things (IoT) and 6^th generation (6 G) communication systems. It is suggested that the power management problems associated with such sensor arrays could be partially met by using self-powering solutions such as triboelectric nanogenerators (TENGs).¹ These power management issues include installation, maintenance and proper management of sensor units, to be used in cutting-edge fields including intelligent sports, tactile systems, health care, security, and document management systems.² Among different sensors that are used to convert information from the physical world to digital format, capacitive sensors are more popular. They are ideal for facilitating human-machine interactions. Numerous investigations are being made to transform capacitive touch sensors to be energy-autonomous. Triboelectric tactile sensors utilize their own energy for sensing via the triboelectric effect, which transforms mechanical impulses into electrical signals. Furthermore, the sensor's self-generated output in response to TENG deformation simplifies the construction of signal processing circuits. Therefore, the implementation of touch sensors using TENG techniques could satisfy the requirement for standalone operation without the need for an additional power supply. Thus, their inclusion in upcoming lifestyle technologies is inevitable.^3,4

The application of TENG-based active sensors for human-machine interactions has been reported in the literature.⁵ For example, Pu et al. reported a super-stretchable transparent tactile sensor system for touch or pressure sensing, which found applications in soft robotics, functional displays, electronic skins, etc.⁶ This sensor system was developed by hybridizing two elastomers as electrification layers while ionic materials such as polyacrylamide-based hydrogels containing lithium chloride were used as the electrode layer. This sandwich-structured sensor system in single electrode mode could sense even a very low input pressure. A soft TENG-based triboelectric tactile sensor array system was also disclosed by Li et al. in 2020, which was made possible by direct ink-writing technology. In terms of voltage modulation, this array of sensors could recognize the tactile outline of physical objects. E-skin for soft robotics could find promising applications by using such printed arrays.⁷ Some other research studies have also shown that flexible and ultrathin TENGs can be used for instantaneous force sensing. These devices could sense force instantly without the need for an additional power supply. They have elaborate applications such as touch detection in intelligent and sustainable electronics. In general, touch sensors open up a plethora of innovative interaction methods in small-scale wearable electronics. TENG-based touch sensors are one step toward self-powered sensors for a sustainable future. Triboelectric touch sensors are highly sensitive to pressure variations because they can detect differentially generated electric impulses from varying applied pressures. The triboelectric tactile sensor's sensitivity has a significant dependence on the materials' selection. Proper combinations of materials provide enough surface area for contact electrification, which in turn improves the triboelectric charge for smaller forces. Greater triboelectric charge generation leads to increased device sensitivity, which produces large and perceptible output signals even at very low input force. Owing to the existence of materials that trap charges in contact layers, a significant quantity of electrical charges will be created, which allows the touch sensor to have a very sensitive feature.

Numerous methods could be used to convert a variety of materials into triboelectric films.

However, developing triboelectric layers with uniform thickness is challenging. Screen printing is considered a very competitive fabrication technology for the quick and scalable production of printed microelectronics.⁸ 2D nano materials are difficult to make large uniform films for device application, but through precise optimization with suitable binder polymers making them suitable candidates for device fabrication. 2D materials are interesting candidates for TENG applications due to their promising physical, chemical, and opto-electronic properties. Hexagonal boron nitride (h-BN) is a layered 2D material, which can be employed in its hybrid/composite form to enhance the output performance of TENGs. With their polymer composite structure, these materials should also act as an electron acceptor to trap electrons. The synthesis methods adopted and the layer thickness of 2D materials do not affect the charge polarity of these materials⁹ Therefore incorporation of triboelectrically negative 2D materials into a polymer composite structure alters the electron affinity properties.¹⁰

Electrospun polymeric nanofibres have wide porosity, good uniformity, large surface area, mechanical flexibility, and stretchability, which offer a new opportunity to act as the triboelectric layer of TENGs and can greatly improve the triboelectric effect and output performance.^11,12 Cellulose-based materials are relatively abundant, renewable, biodegradable, flexible, and mechanically strong. Interestingly, they can easily develop positive charges on their surface through electron loss, which makes them suitable candidates for eco-friendly TENGs.¹³ The triboelectric potential difference between the contact layers is found to be greatly enhanced when a positive cellulose material and a triboelectrically negative material are properly combined in a TENG configuration.^14–16 It is reasonable to expect that a cellulose and h-BN based triboelectric pair can generate a sizable polarity difference.¹⁷

Drawing inspiration from these leads, we have developed a flexible TENG-based self-powered tactile sensor. Screen-printed BNNS polyvinyl pyrrolidone composite ink (BN–PVP) on a BoPET substrate and surface-modified electrospun cellulose acetate (ES-CA) nanofibers served as the two contact materials. This flexible TENG in vertical contact separation (CS) mode produced a massive output voltage of ∼1200 V and short-circuit current density of 1.2 mA m⁻² respectively. Power density reached 1.4 W m⁻². Here, the incorporation of 2D nanosheets in a polymer matrix greatly improves the power output of the fabricated TENG. Furthermore, a self-powered tactile sensor was fabricated by slightly modifying the TENG device structure. Remarkably, the developed TENG-based tactile sensor was able to detect and distinguish a force down to ∼0.05 N. The sensitivity achieved here is 3.98 V N⁻¹ for comparably lower forces of less than 2 N, whereas, for higher forces in the range of 2–10 N, it is found to be 1.84 V N⁻¹.

2. Experimental

2.1. Materials

Pristine h-BN (powder size ∼ 1 μm and 98% purity) was procured from Sigma-Aldrich. The binder polymer polyvinyl pyrrolidone (PVP) and cellulose acetate (CA) of average molecular weight M_n ∼30 000 were also purchased from Sigma-Aldrich (India). Dimethyl formamide (DMF), a dispersing organic solvent, is acquired from HPLC (India). The substrate material Mylar® (BoPET) was received from DuPont, USA. Ethanol for printing was procured from Merck Chemicals, India. All reagents and chemicals were used without further purification.

2.2. BN–PVP ink formulation

Liquid phase exfoliation of bulk h-BN in an organic solvent, DMF, produced 2D BNNSs (detailed characterization in ESI S1 and S2†). Through a solution mixing process, these BNNSs were combined with the polymer binder PVP (1 : 1 weight ratio) dispersed in 1 mL of organic solvent taken in a culture tube. Later, the highly viscous dispersion was magnetically stirred for 12 hours at 700 rpm rotation speed, and continuous magnetic stirring converted the mixture into an ink (BN–PVP) form depicted in Fig. 1(a). In this method, PVP plays a dual role as a polymer matrix for the composite and a binder for the formulated ink. A similar procedure was also adopted for pure polymer dispersion without BNNSs.

Fig. 1 (a) Schematic depicting the BN–PVP ink formulation; (b) screen printing on BoPET. (c) Schematic illustration representing the electrospinning process for developing a cellulose acetate nanofiber mat on aluminum foil.

2.3. Printing of the triboelectric layer using BN–PVP ink

Highly viscous BN–PVP ink was screen-printed on a flexible substrate (BoPET) as shown in Fig. 1(b). Here a home-made nylon screen with predesigned square patterns was used for printing the composite ink. As an initial stage, the rheology-controlled BN–PVP ink was spread over the nylon screen. Then, the rubber squeegee motion horizontally over the screen surface pushes the ink through the open pores of the screen. Consequently, the desired pattern was inscribed on the substrate surface that was placed under the screen at a specific snap-off distance. The number of printing strokes can be varied to control the thickness of the printed film. The smooth-printed BN–PVP film on the BoPET substrate was then dried at 60 °C in an oven. This printed BN–PVP ink on BoPET acts as the negative layer of the fabricated TENG. Through a similar printing procedure, a pure polymer PVP film (without the BNNS filler) on the BoPET substrate was also made to compare the device outputs for elucidating the role of BNNS fillers.

2.4. Preparation of an electrospun-cellulose acetate (ES-CA) nanofiber mat

Electrospinning was used to generate cellulose acetate nanofibers. For this purpose, we used a well-dispersed homogeneous solution of CA in a mixture of acetone and DMF solvent. Then, the solution was loaded in a syringe of the spinner that has a ∼1 mm inner needle diameter. The CA solution's flow rate was set around 1 mL h⁻¹ at a spinning voltage of 15 kV, to produce nanometer dimensioned fibers. During the electrospinning operation, solution was continuously extruded towards the rotating collector drum covered with aluminum foil (Al), which is kept 15 cm away from the nozzle tip. The spinning collector drum (rotation speed ∼900 rpm) and needle tip were linked to the ground (negative) and positive electrodes, respectively. The CA nanofiber mat developed on the surface of aluminum foil was dried in a vacuum oven at 60 °C overnight.^18,19 A schematic illustration of the electrospinning procedure for producing cellulose acetate nanofibers over the surface of aluminum foil is shown in Fig. 1(c).

2.5. BN–PVP/ES-CA TENG-based tactile sensor fabrication

The flexible BN–PVP/ES-CA TENG was designed in contact separation (CS) mode using a flexible BoPET substrate. The photograph and the schematic of the fabricated TENG device are presented in Fig. 2(a) and (b). BN–PVP ink printed on the BoPET substrate and the electrospun-cellulose acetate nanofiber mat were taken as the two opposite contact materials for the TENG. The polymer substrate, BoPET was employed as the support structure for assembling positive and negative TENG contact materials. The negative BN–PVP on the BoPET with the back copper electrode layer and positive ES-CA layer on Al were arranged on the rectangular flexible frame using suitable adhesives. The copper (Cu) electrode layer was arranged as a sandwich layer between the rectangular frame and negative triboelectric material whereas aluminium was arranged as a sandwich layer between the rectangular frame and positive triboelectric material. Electrodes were connected to the measuring instruments through copper tape and wires. Finally, the output performance of the fabricated device was evaluated with the aid of a custom-made cyclic force impactor system. Furthermore, the geometry of the BN–PVP/ES-CA TENG was tweaked to fabricate a tactile sensor by modifying its contact area and separation distance between the contact materials. Two contact materials of 1 cm² area were positioned 80 μm apart from one another, while the spacing was maintained using a BoPET layer. The entire system was encapsulated within a protective layer using the vacuum sealing method. Fig. 2(a) presents the photograph of the BN–PVP/ES-CA TENG and Fig. 2(b) presents the schematic of the device and Fig. 2(c) presents a schematic representing the cross-sectional view of the tactile sensor.

Fig. 2 (a) Photograph of the BN–PVP/ES-CA TENG. (b) Schematic model depicting the BN–PVP/ES-CA TENG. (c) Schematic representation of the tactile sensor (cross-sectional view).

2.6. Characterization methods

A comprehensive analysis of the developed BN–PVP ink, its coating on the polymer substrate, ES-CA nanofibers, BN–PVP/ES-CA TENG device, and also the BN–PVP/ES-CA TENG-based tactile sensor was carried out using various characterization tools. Rheological characteristics such as viscosity and visco-elastic properties of the BN–PVP composite inks were tested using a rheometer (Rheo Plus 32, Anton Paar, Graz, Austria). This study examined the developed ink viscosity at varying shear rates from 0 to 100 s⁻¹. The FT-IR spectra of the BN–PVP ink powder and pure polymer PVP were recorded using an FTIR instrument (Nicolet Magna 560, Thermo Scientific, Massachusetts, USA). The contact angle measurement using the sessile drop method (DSA 30, KRUSS GmbH Hamburg, Germany) was used to investigate the wettability of ink on the substrate surface. Wide-angle X-ray diffraction (WAXD) measurements of the cellulose acetate nanofibers were performed using a Xeuss WAXS system (Xenocs, Grenoble, France). The surface of screen-printed BN–PVP films on a polymer substrate was investigated by the scanning electron microscopy technique (Zeiss EVO 50, Oberkochen, Germany) at an accelerating voltage of 15 kV. A polarising optical microscope with a CCD camera (DM2700P, Leica, Germany) was used to capture the optical images of the printed films. All the electrical measurements of the fabricated PVP/ES-CA TENG with and without BNNSs were carried out using a Keithley 2450 source measurement unit system (Tektronix, USA). The device's short-circuit current was measured using a Stanford Research low-noise SR570 current preamplifier setup. A custom-made modified force impactor system was used as the mechanical input source for the operation of the TENG, which can impart ∼10 N force at variable frequencies. Many custom-made resistor banks and capacitor banks were used to study the load-dependent output power density and charging profile of the developed TENGs. All the device characterization studies were conducted under ambient conditions. A motorized force impactor with a precision force gauge (Mark-10, ESM303 Stand and M5-50 Gauge) was used to supply the input forces for analyzing the performance of the fabricated touch sensor.

3. Results and discussion

3.1. Rheology and structural analysis of BN–PVP ink

Precise control over the ink composition is necessary for the printing to be effective and to provide the desired qualities for the formulated inks. The even distribution of filler particles (BNNSs) in the vehicle system controls the ink's colloidal stability. The well dispersed filler in DMF solvent with suitable bio-compatible binder material, PVP, stabilizes the filler nanosheets in dispersion and modifies the rheology appropriate for screen printing.²⁰ The adherence of the ink to the substrate surface was also improved using PVP. The flow characteristics of the BN–PVP composite ink were studied to investigate the printability of the prepared ink. As shown in Fig. 3(a), the viscosity variation of the BN–PVP ink shows typical shear thinning behavior with the variation of the shear rate from lower 1 s⁻¹ to higher 100 s⁻¹. Given in the inset of Fig. 3(a) is the photographic representation of the BN–PVP ink. This colloidal ink mixture displays an ideal pseudoplastic behaviour. In other words, the apparent viscosity decreased linearly as a function of the shear rate. At a shear rate of 10 s⁻¹, it shows an ideal viscosity of about 3.2 Pa s which is within the required range suitable for screen printing.²¹ The viscoelastic properties of the BN–PVP ink were further analyzed using the diagrams of the storage modulus (G′) and loss modulus (G′′), as depicted in Fig. 3(b). In the entire strain region, the storage modulus shows a decreasing trend compared to the loss modulus values (G′′ > G′). This implies that the BN–PVP ink possesses an ideal viscous liquid-like behaviour rather than a solid-like nature that promotes solidification of the inks during printing.²²

Fig. 3 (a) Viscosity-shear rate study of the BN–PVP ink. The inset of (a) presents the photograph of the BN–PVP ink. (b) Storage modulus and loss modulus variation of BN–PVP ink with respect to shear strain. (c) Comparison of FTIR spectra of PVP and BN–PVP and (d) the contact angle of BN–PVP ink on the BoPET substrate.

A comparison of FT-IR spectra of the pure polymer and the composite BN–PVP is displayed in Fig. 3(c). A distinctive peak was seen at 1652 cm⁻¹ in pure polymer PVP, corresponding to the C–O stretching vibrations. This recorded peak at 1652 cm⁻¹ is also evident in the FT-IR spectra of the BN–PVP composite. The other peaks of the pure polymer centered at 1423 cm⁻¹ represent bending of the C–H bond, while the peak at 1288 cm⁻¹ shows wagging vibrations of CH₂.²³ These polymer peaks are absent in the BN–PVP sample because these vibrational lines and the peak for B–N in-plane stretching vibrations of the BNNSs overlap in this window. Two more peaks are seen in the BN–PVP composite in addition to the polymer peak at 1652 cm⁻¹. These peaks are located at 760 cm⁻¹ and 1342 cm⁻¹ representing the out of plane B–N–B bending vibration and B–N stretching vibration respectively.²⁴

The contact angle measurement of the formulated BN–PVP ink was used to determine its wetting behaviour on the substrate surface and its interface characteristics. A static ink droplet on the BoPET surface with an average contact angle, θ = 64.9°, is shown in Fig. 3(d). This contact angle formed on the substrate surface might provide vital clues about its wetting nature. Lower contact angle values, almost equal to θ = 0°, signify a more liquid spreading nature of the substrate surface. In other words, the substrate surface is more hydrophilic and has better wetting when θ is less than 90°. On the other hand, when θ is more than 90°, it is commonly referred to as hydrophobic and has a poor wetting quality.^25,26 Better wettability and affinity of the BN–PVP ink on the BoPET substrate are shown by the recorded contact angle, which is 64.9°.

3.2. Morphology analysis of BN–PVP on BoPET

The representative SEM micrograph of the surface of screen-printed BN–PVP coating on BoPET is exhibited in Fig. 4(a). The films produced by the screen printing process are homogeneous and have excellent ink distribution across the substrate. The PVP polymer matrix contains the exfoliated BNNSs embedded in it with minimal aggregation. The optical image of the BN–PVP ink surface is also displayed in Fig. 4(b). The thickness evaluation of the BN–PVP coating on the substrate surface is shown in Fig. 4(c) and (d). The average thickness of the films was measured to be approximately 9 μm, where the film was formed after two screen printing cycles. Factors such as solid content contribution, viscosity of the ink and the number of printing strokes determine the thickness of the printed film.^27,28

Fig. 4 (a) SEM image of BN–PVP ink surface on BoPET. (b) Optical image of the BN–PVP ink surface on BoPET. (c) & (d) SEM cross-sectional images of BN–PVP ink on the BoPET substrate showing the film's thickness.

3.3. Structure and morphology of cellulose acetate nanofibers

Sustainable cellulose materials are regarded as triboelectrically positive among the tested triboelectric materials.²⁹ The cellulose fiber-based triboelectric positive layer is a direct consequence of several procedural steps, including electrostatic spinning, laser treatment, and plasma treatment. These strategies may enhance the functionality of TENGs and operate more effectively.³⁰ As a positive TENG material, ES-CA nanofibers were developed in this study. The FT-IR spectra of ES-CA nanofibers in Fig. 5(a) show strong characteristic absorption peaks at 1751 cm⁻¹ and 1240 cm⁻¹. These peaks arise as a result of C=O (carbonyl) stretching vibrations and acetyl C–O stretching respectively. The additional peak observed at 1052 cm⁻¹ is also due to the C–O stretching vibrations.¹⁸

Fig. 5 (a) FT-IR spectra and (b) WAXD pattern of ES-CA nanofibers. (c) SEM micrograph of ES-CA nanofibers and (d) optical image of ES-CA nanofibers on Al foil.

The less sharp, comparatively weak intensity peaks in the WAXD pattern of ES-CA nanofibers are presented in Fig. 5(b). This suggests that the structure of the material is amorphous and lacks a regular molecular structure.³¹ The SEM micrograph of the prepared cellulose acetate nanofibers from the electrospinning method is included in Fig. 5(c). From these images, it can be seen that thin nanofibers with varying widths and lengths do not form any beads. Its smooth surfaces, which indicate the purity of the fibers generated, may be single or joined to many sections. The optical image of the ES-CA nanofibers on the Al substrate is presented in Fig. 5(d).

3.4. Characteristics of the BN–PVP/ES-CA TENG

The BN–PVP/ES-CA TENG was developed using two triboelectrically opposite contact materials, ES-CA nanofibers as the positive triboelectric layer and the BNNS embedded PVP polymer as the opposite contact layer, having a negative tribopolarity. ES-CA nanofibers with higher specific surface area and porosity could provide better mechanical contact during the device operation, which enhanced the magnitude of the induced triboelectric charges.³² In addition to this, the charge trapping nature resulting from the inclusion of BNNSs as the filler material for the polymer-based ink further improved the TENG output by confining more charges as a result of contact electrification on the surface. The presence of 2D BNNSs in the BN–PVP composite structure also transformed the surface charge potential into a larger negative value, which resulted in a significant triboelectric potential difference between the two top and bottom contact layers.^33,34

Fig. 6 schematically illustrates the basic working mechanism of the BN–PVP/ES-CA TENG. This device uses an ES-CA nanofiber based nonwoven mat as the positive layer and BN–PVP ink on the BoPET substrate as the negative layer, which are arranged in a CS mode configuration. The forced contact between these layers generates a collection of equally dense opposite charges on the layer surface. When the applied force is removed, the separation between triboelectric layers creates an electrical potential difference, to maintain an equilibrium. It drives the flow of electrons from the top electrode to the bottom one owing to electrostatic induction. As a result, an instantaneous current flow is observed in the external circuit. The external circuit allows the flow of electrons to move back and forth, due to the sequential operation of pressing and releasing resulting in an alternating electrical output.^19,35

Fig. 6 Schematic representation of the working of the flexible BN–PVP/ES-CA TENG in CS mode configuration. (a) Initial contact stage, (b) separation stage, (c) separated stage of saturation, and (d) further contact stage.

The contact area of the triboelectric materials was found to be influencing the output performance of the BN–PVP/ES-CA TENG. This was investigated by systematically fabricating TENGs with triboelectric layers of constant separation gaps but varying contact areas of 1, 4, and 9 cm² respectively. Proportionate variation is observed in output voltage and short circuit current density as a function of the device's contact area are depicted in Fig. 7(a) and (b). As expected, enhanced charge transfer and surface charges are responsible for the observed improvement in the device performance.^36,37 For example, the voltages of devices with contact areas of 1, 4, and 9 cm² are 110 V, 520 V, and 1200 V respectively. The corresponding current densities are 0.27 mA m⁻², 0.8 mA m⁻², and 1.2 mA m⁻². Similarly, to examine the influence of the separation gap on the device output, a series of TENGs were built with a constant contact area of 9 cm² with various separation gaps (0.5 cm, 0.7 cm, and 1.0 cm). Fig. 7(c) and (d) illustrate the fluctuation in output voltage and short-circuit current density of these BN–PVP/ES-CA TENGs with various separation gaps. Both the voltage and current density show an increasing trend with the separation gap also, and the maximum voltage of the device (∼1200 V) was achieved with a 10 mm separation between the contact layers.

Fig. 7 (a) & (b) Output voltage & current density of the BN–PVP/ES-CA TENG with varying contact areas. (c) & (d) Output voltage & current density of the BN–PVP/ES-CA TENG with different separation gaps.

In order to examine the mechanical energy harvesting performance of the BN–PVP/ES-CA TENG, a specifically modified force impactor setup was used. The electrical measurements of the printed TENG with a 9 cm² contact area and 10 mm separation gap were performed. The output voltage and current density comparison plots are shown in Fig. 8(a) and (c), which exhibit an improved triboelectric performance by the introduction of BNNSs. The charge-trapping nature of 2D nanomaterials in its composite structure greatly contributes to the surface charge density of a nanocomposite based TENG. In addition, the multi-layered device construction with an extra charge storage layer such as PET between the frictional tribolayer and the electrode in a composite multi-layer structure also increased the output compared to the single-layer structure by effectively storing the generated charges.¹¹ The BN–PVP on BoPET and ES-CA nanofibers employed in the flexible screen-printed TENG has produced a whopping output voltage of 1200 V and a current density of 1.2 mA m⁻², respectively. This fabulous result is way better than that of the PVP/ES-CA TENG without BNNSs. This is because the PVP/ES-CA TENG could only yield an output voltage of 180 V and a current density of 0.23 mA m⁻² respectively.

Fig. 8 (a) Output voltage comparison of PVP/ES-CA & BN–PVP/ES-CA TENGs having device ∼ area 9 cm². (b) Power density of the PVP/ES-CA TENG without BNNSs in its polymer matrix. (c) Current density comparison of PVP/ES-CA & BN–PVP/ES-CA TENGs. (d) Power density of the BN–PVP/ES-CA TENG.

From the impedance matching experiments, the power densities of both the PVP/ES-CA TENG and BN–PVP/ES-CA TENG were investigated by using different resistors with varying resistance values and are shown in Fig. 8(b) and (d) respectively. With increasing resistance, the BN–PVP/ES-CA TENG's output voltage progressively increases to a maximum of 1200 V. As evident from the load-matching study, the output power initially increases and then decreases. The calculated instantaneous power density of the BN–PVP/ES-CA TENG on an external load reaches a maximum of 1.4 W m⁻² at 200 MΩ resistance and is shown in Fig. 8(d). This observed power density is almost 100 times higher than the power density of the PVP/ES-CA TENG (0.015 W m⁻²), which is shown in Fig. 8(c). It is obvious that small-scale electronic devices can be directly powered using this high output TENG.

Fig. 9(a) depicts the capacitor charging curves via a bridge rectifier using the BN–PVP/ES-CA TENG, utilizing various capacitors with capacitance values of 0.5 μF, 1 μF, 5 μF, and 10 μF respectively. For capacitors with higher capacitance values, longer charging times are needed. The BN–PVP/ES-CA TENG could charge the above capacitors to an adequate level which is good enough to power handheld electronic devices and sensor units. Fig. 9(b) depicts the relationship between TENG's output characteristic and the frequency of input mechanical force. It shows significant enhancement in the output voltage as the frequency increases from lower to higher values. Since the maximum output was recorded at 5 Hz frequency, all other subsequent measurements of the BN–PVP/ES-CA TENG were conducted under 5 Hz frequency. The reliability analysis of the BN–PVP/ES-CA TENG was further conducted by more cycles of operation. Even after 12 000 cycles of contact and separation movement, a stable output without any degradation is maintained by the developed TENG device. The reliability analysis of the BN–PVP/ES-CA TENG is presented in Fig. 9(c).

Fig. 9 (a) BN–PVP/ES-CA TENG capacitor charging profile with different capacitors of varying capacitances. (b) Output voltage analysis of the BN–PVP/ES-CA TENG with respect to frequency. (c) Reliability study of the BN–PVP/ES-CA TENG for 12 000 cycles of device operation.

Fig. 10(a) and (b) present the variation of output voltage and short-circuit current density of the BN–PVP/ES-CA TENG at diverse magnitudes of biomechanical forces. These results show a significant increment in the intensity of the applied forces. The impact of force is obviously higher for palm pressing compared to single-finger tapping and multiple-finger tapping. Correspondingly, the output results are also considerably higher for palm pressing. Relatively inferior results were obtained for single-finger pressing. A higher impact force can provide better surface contact between the triboelectric materials and the output will be higher.³⁵ As shown, an output voltage of 1640 V and a short-circuit current density of 2 mA m⁻² were obtained under the palm pressing condition which are well above that for single-finger (500 V and 0.66 mA m⁻²) and multiple-finger tapping (1300 V and 1.4 mA m⁻²).

Fig. 10 (a) & (b) Output voltage and current density of the BN–PVP/ES-CA TENG under various biomechanical pressures: single finger, several fingers, and palm pressing.

3.5. BN–PVP/ES-CA TENG as a self-powered tactile sensor

The development of devices with haptic perception is essential for future human-machine interface applications. Self-powered flexible tactile sensors based on the triboelectric effect could convert tactile data directly into electrical signals and are widely being used in robotics and prosthetic applications.^36–38 The BN–PVP/ES-CA TENG's remarkable potential as a self-powered tactile sensor for touch-sensing application is already demonstrated by the force-sensitivity study conducted on it. In order to investigate application of the fabricated device for pressure sensing, the BN–PVP/ES-CA TENG's structural configuration was suitably modified by minimizing its device parameters. The device contact area was reduced to 1 cm² and the separation gap adjusted to approximately 85 μm. The fabricated self-powered tactile sensor is displayed in the inset image of Fig. 11(a) and the experimental setup is shown in Fig. 11(c).

Fig. 11 Tactile sensing performance of the BN–PVP/ES-CA TENG based tactile sensor. (a) Output voltage variation with respect to applied force ranging from 0.05 N to 10 N. (b) The sensitivity of the tactile sensor in the force range 0.05 N to 10 N. (c) Photograph of the motorized force impactor set-up with a precision force gauge used for studying tactile sensing performance of the device.

To examine the sensitivity of the tactile sensor, different magnitudes of forces ranging from 0.05 N to 10 N with slow movement speeds have been applied to the device. It is discovered that the output voltage of the BN–PVP/ES-CA TENG based tactile sensor is clearly force-sensitive. Fig. 11(a) indicates the output voltage variation produced by the tactile sensor under different pressure conditions. These results indicate the triboelectric sensor's linear dependence on the applied force and also its ability to detect very minute forces. This linear trend can be explained by the more effective surface contact between the triboelectric materials under higher force conditions, which improved its output results.^39,40

The observed resolution is ∼0.05 N for the triboelectric tactile sensor in the current experimental set-up. As the force changes from 0.05 N to 2 N, the tactile sensor's output voltage varies linearly from millivolts to 9 V. Fig. 11(b) illustrates the corresponding sensitivity in this lower force region, which is ∼3.98 V N⁻¹. In the higher force range from 2 N to 10 N, a similar trend can be seen. At 10 N force, the maximum voltage could reach up to 23 V, and the observed sensitivity is 1.84 V N⁻¹, which is shown in Fig. 11(b). The tactile sensor's sensitivity is higher in the lower force regime because there is a greater change in the generation of triboelectric charges. As the force increases, the device reaches a saturation point and sensitivity decreases thereafter. Hence, the developed BN–PVP/ES-CA TENG based tactile sensor is ideal to detect even the smallest force variations with exceptional sensitivity,^41,42 which can be readily implemented in a multitude of application fields, such as robotics and prosthetics, for achieving tactile perceptions precisely.^43,44

4. Conclusions

In this work, we demonstrate the design and realization of a flexible TENG for kinetic energy harvesting and also for self-powered tactile sensing. Electrospun cellulose acetate (ES-CA) nanofibers and screen-printed BNNS composite ink (BN–PVP) on a BoPET substrate were used as the positive and negative triboactive materials respectively. This flexible TENG design with a 10 mm separation gap and 9 cm² contact area yielded an impressive result of output voltage ∼1200 V and 1.2 mA m⁻² short-circuit current density. The resultant peak power density could go up to 1.4 W m⁻². This device performance is comparably much higher than that of a similar PVP/ES-CA TENG fabricated using PVP alone (without 2D nanosheets). Quite significantly, the power density of the BN–PVP/ES-CA TENG is over 100 times better than that of the reference device without 2D fillers (PVP/ES-CA TENG). On top of that, the new device is flexible and durable for long-term operations, making it a more suitable practical choice for powering wearable and portable devices.

Through suitable choice of materials, the high power delivering capability of the TENG could be transformed into a self-powered energy autonomous tactile sensor. For this, a BN–PVP/ES-CA based touch sensor was fabricated, for detecting input forces of very low magnitude (0.05 N). Remarkably, the low force sensitivity of the sensor is 3.98 V N⁻¹, whereas, for comparably higher forces (2–10 N), it is found to be 1.843 V N⁻¹. In brief, the innovative material design and device architecture lead to the development of a flexible high energy TENG and tactile sensor, which offers a viable solution towards energy autonomous wearable devices for IoT applications.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

Author A. S. Bhavya is thankful to the Council of Scientific and Industrial Research (CSIR), India for the award of a research fellowship. Authors, K. P. Surendran and Achu Chandran are pleased to acknowledge the Council of Scientific and Industrial Research (CSIR), New Delhi, India for financial support to carry out this work, through a Mission Mode Project [MMP 035201] & a Fast Track Translational Project [FTT040506]. The authors further acknowledge the funding support from Science & Engineering Research Board (SERB), New Delhi, India, under Core Research Grant projects [CRG/2023/004068] & [CRG/2021/00495], Department of Science and Technology (DST), New Delhi, India under Device Development Project [DST/TDT/DDP-07/2021] & Advanced Manufacturing Technology Project [DST/TDT/AMT/2021/001G], and also Department of Space, Bengaluru through two ISRO Respond Projects [ISRO/RES/3/884/21-22] & [SRO/RES/3/1007/24-25]. Finally, the authors thankfully acknowledge various scientists-in-charge of different instruments and testing supports available in CSIR-NIIST.

References

1. Y. Zhou, M. Shen, X. Cui, Y. Shao, L. Li and Y. Zhang, Nano Energy, 2021, 84, 105887.
2. Q. Zhou, J. Pan, S. Deng, F. Xia and T. Kim, Adv. Mater., 2021, 35, 2008276.
3. Y. Quan, E. Wang, H. Ouyang, L. Xu, L. Jiang, L. Teng and J. Li, Adv. Sci., 2024, 1(11), 2409914.
4. S. Zuk, A. Pietrikova and I. Vehec, Microelectron. Int., 2018, 3, 153–157.
5. H. Ouyang, Z. Li, M. Gu, Y. Hu, L. Xu, D. Jiang, S. Cheng, Y. Zou, Y. Deng, B. Shi, W. Hua, Y. Fan, Z. Li and Z. Wang, Adv. Mater., 2021, 1(9), 2102302.
6. X. Pu, M. Liu, X. Chen, J. Sun, C. Du, Y. Zhang, J. Zhai, W. Hu and Z. L. Wang, Sci. Adv., 2017, 3, 1–11.
7. H. Li, X. Fang, R. Li, B. Liu, H. Tang, X. Ding and Y. Xie, Nano Energy, 2020, 78, 105288.
8. Y. Xu, M. G. Schwab, A. J. Strudwick and I. Hennig, Adv. Energy Mater., 2013, 8, 1035–1040.
9. M. Seol, S. Kim, Y. Cho, E. K. Byun, H. Kim, J. Kim, S. K. Kim, S. W. Kim, H. J. Shin and S. Park, Adv. Mater., 2018, 30, 1801210.
10. H. Kuang, Y. Li, S. Huang, L. Shi, Z. Zhou, C. Gao, X. Zeng, R. Pandey, X. Wang, S. Dong and X. Chen, Nano Energy, 2021, 80, 105561.
11. B. U. Ye, B. J. Kim, J. Ryu, J. Y. Lee, J. M. Baik and K. Hong, Nanoscale, 2015, 7(39), 16189–16194.
12. C. Li, Y. Yin, B. Wang, T. Zhou, J. Wang, J. Luo, W. Tang, R. Cao, Z. Yuan, N. Li and X. Du, ACS Nano, 2017, 11, 10439–10445.
13. S. W. Chen, X. Cao, N. Wang, L. Ma, R. H. Zhu, M. Willander, Y. Jie and Z. L. Wang, Adv. Energy Mater., 2016, 7, 1601255.
14. Z. Niu, W. Cheng, M. Cao, D. Wang, Q. Wang, J. Han, Y. Long and G. Han, Nano Energy, 2021, 87, 106175.
15. K. Shi, H. Zou, B. Sun, P. Jiang, J. He and X. Huang, Adv. Funct. Mater., 2020, 30(4), 1904536.
16. Y. Liu, J. Ping and Y. Ying, Adv. Funct. Mater., 2021, 1(16), 2009994.
17. G. Khandelwal, N. P. Maria Joseph Raj and S. J. Kim, Adv. Energy Mater., 2021, 11(33), 1–32.
18. J. Zhou, H. Wang, C. Du, D. Zhang, H. Lin, Y. Chen and J. Xiong, Adv. Energy Sustainability Res., 2022, 3(5), 2100161.
19. Y. Zheng, L. Cheng, M. Yuan, Z. Wang and L. Zhang, Nanoscale, 2014, 6(14), 7842–7846.
20. D. W. Kim, J. H. Lee, J. K. Kim and U. Jeong, NPG Asia Mater., 2020, 12(1), 6.
21. N. Joseph, J. Varghese and M. T. Sebastian, J. Mater. Chem. C, 2016, 4, 999–1008.
22. X. Tang, K. Wu, X. Qi, H. Kwon, R. Wang, Z. Li, H. Ye, J. Hong, H. H. Choi, H. Kong, N. Lee, S. Lim, Y. J. Jeong and S. Kim, ACS Appl. Nano Mater., 2022, 5(4), 4801–4811.
23. H. Li, X. Li, J. Liang and Y. Chen, Adv. Energy Mater., 2019, 1(3), 1803987.
24. E. J. J. Samuel, P. Sathiyaraj, M. A. Deminskii and K. M. Konina, Mater. Sci. Eng., 2017, 1, 012043.
25. M. I. J. Hassan, M. I. J. Haider, A. U. Hamid, I. S. M. Ikram and A. R. U. Qumar, Appl. Nanosci., 2020, 10, 3525–3528.
26. M. Gru, D. G. Waugh, J. Lawrence, N. Langer and D. Scholz, Langmuir, 2019, 35(38), 12356–12365.
27. G. D. Luca and E. Amendola, Open Chem. J., 2012, 11(1), 35–45.
28. M. Franco, A. Motealleh, C. M. Costa, L. Hilliou, N. Perinka, C. Ribeiro, J. C. Viana and P. Costa, Adv. Eng. Mater., 2022, 1(11), 2101258.
29. J. Ju, Z. Zhao, X. Du, X. Sun, J. Liu and W. Li, Ceram. Int., 2023, 49(17), 28854–28863.
30. Y. Song, Z. Shi, G. H. Hu, C. Xiong, A. Isogai and Q. Yang, J. Mater. Chem. A, 2021, 9(4), 1910–1937.
31. F. Rahmawati, I. Fadillah and M. Mudjijono, J. Mater. Environ. Sci., 2017, 8(1), 289–297.
32. H. Varghese, H. M. A. Hakkeem, K. Chauhan, E. Thouti, S. Pillai and A. Chandran, Nano Energy, 2022, 98, 107339.
33. S. Cheon, H. Kang, H. Kim, Y. Son, J. Y. Lee, H. Shin, S. Kim and J. H. Cho, Adv. Funct. Mater., 2018, 1(7), 1703778.
34. A. S. Bhavya, H. Varghese, A. Chandran and K. P. Surendran, Nano Energy, 2021, 90, 106628.
35. B. Bera and K. Wang, Imp. J. Interdiscip. Res., 2016, 2(10), 1263–1271.
36. S. Nuthalapati, A. Chakraborthy, I. Arief, K. K. Meena, K. Ruthvik, R. R. Kumar, K. U. Kumar, A. Das, M. E. Altinsoy and A. Nag, IEEE Journal on Flexible Electronics, 2024, 3, 393–400.
37. N. Madathil, S. Potu, J. Pani, L. Bochu, A. Babu, H. Borkar, P. Kodali, U. K. Khanapuram and R. K. Rajaboina, ACS Appl. Electron. Mater., 2024, 6(8), 5563–5574.
38. R. K. Rajaboina, U. K. Khanapuram and A. Kulandaivel, Adv. Sensor Res., 2024, 3(10), 2400045.
39. H. Varghese, H. M. Abdul, M. Farman, E. Thouti, S. Pillai and A. Chandran, Results Eng., 2022, 16, 100550.
40. J. Syamini and A. Chandran, ACS Appl. Electron. Mater., 2023, 5(2), 1002–1012.
41. Y. H. Ko, G. Nagaraju, S. H. Lee and J. S. Yu, ACS Appl. Mater. Interfaces, 2014, 6(9), 6631–6637.
42. H. Varghese and A. Chandran, ACS Appl. Mater. Interfaces, 2021, 13(43), 51132–51140.
43. L. Lu, G. Hu, J. Liu and B. Yang, Adv. Sci., 2024, 1(9), 2309894.
44. J. Li, C. Carlos, H. Zhou, J. Sui, Y. Wang, Z. Silva-pedraza, F. Yang, Y. Dong, Z. Zhang, T. A. Liu, B. Mao, Y. Hacker and X. Wang, Nat. Commun., 2020, 1(11), 2023.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d5se00302d

---