Electrospinning Machine| Enhancing output performance of triboelectric nanogenerator based on high-dielectric Ti3CNTx/PVA nanofiber membrane for human–computer interaction applications

With the iterative advancement of human-computer interaction (HCI) technology, its core carriers have extended from traditional keyboards and mice to touchscreens, voice interaction, and gesture recognition systems. Particularly in VR/AR and metaverse scenarios, the demand for wearable devices with high sensitivity, comfort, and precise recognition is increasingly prominent. However, current TENG-based gesture recognition technologies primarily rely on gloves or finger-worn devices, which suffer from drawbacks such as poor breathability, significant response delays, and low recognition accuracy. Traditional designs mostly adopt glove- or finger-mounted configurations, but these often compromise wearability and long-term user comfort. Therefore, developing flexible triboelectric materials with ultrathin structures, high breathability, and intelligent signal processing capabilities to achieve precise motion capture via skin-conformal contact has become a key path to break through current technical barriers.

Recently, Associate Researcher Huamin Chen from the Institute of Semiconductors, Chinese Academy of Sciences, published the latest research in Chemical Engineering Journal titled "Enhancing output performance of triboelectric nanogenerator based on high-dielectric Ti₃CNTₓ/PVA nanofiber membrane for human-computer interaction applications." The first authors of the paper are Associate Professor Ruilai Liu from Wuyi University, Master's students Haihang Feng and Chaoyang Sun from Fujian University of Technology, and the corresponding authors are Associate Researcher Huamin Chen, Researcher Mingcen Weng, and Associate Professor Yu Xiao.

The research team successfully prepared a Ti₃CNTₓ/PVA nanofiber membrane via an optimized electrospinning process. The innovation lies in uniformly embedding high-dielectric Ti₃CNTₓ nanosheets into the PVA matrix, significantly enhancing the material's mechanical properties. This nanofiber membrane not only combines high breathability, good biocompatibility, and excellent mechanical flexibility but also greatly boosts the output performance of the triboelectric nanogenerator (TENG) due to its high dielectric constant. Additionally, the team broke through traditional glove/finger-integrated models by innovatively developing a thin-film-structured MP-TENG sensing system. The device conformally adheres to the backhand skin via a contact design, capturing gesture motion features while avoiding movement interference. Combined with a self-developed multilayer perceptron (MLP) neural network, the system achieved 96.36% recognition accuracy on an international standard gesture dataset. This technological breakthrough not only provides new optimization ideas for TENG interface engineering but also opens a flexible electronics integration paradigm for wearable HCI development.

Figure 1: Preparation and structural characterization of Ti₃CNTₓ/PVA hybrid nanofiber membranes.

As shown in Figure 1, the researchers prepared an MP nanofiber membrane with high breathability and flexibility via electrospinning. SEM images reveal that MXene/PVA forms a 3D interlocked network with uniform fiber diameter (Figures 1c-d). Its smooth surface and homogeneous element distribution (EDS confirms uniform dispersion of C/N/O/Ti atoms; XRD shows characteristic peaks at 6.24° and 19.26°) synergistically enhance interfacial contact efficiency. FTIR analysis confirms that MXene-PVA hydrogen bonding strengthens the molecular network (Figures 1e-g). The material exhibits a tensile strength of 6.9 MPa (431% higher than pure PVA) and 98.7% fracture strain (Figure 2h), attributed to hydrogen bonds inhibiting molecular chain slippage. This heterostructure design and exceptional natural hydrolysis properties provide a new paradigm for degradable electronic devices.

Figure 2: Dielectric properties of MP nanofiber membranes and XPS spectra of Ti₃CNTₓ and Ti₃C₂Tₓ samples.

As shown in Figure 2, the study prepared high-dielectric triboelectric materials by doping MXene nanosheets into a PVA matrix. The -F/-OH groups enriched on MXene surfaces greatly enhance the material's electronegativity, with Maxwell-Wagner-Sillars interfacial polarization between MXene and PVA being the core mechanism for performance breakthrough. Experiments show that increasing MXene concentration significantly raises the composite's dielectric constant, with the 5 mg/mL sample exhibiting optimal dielectric performance and polarization capability. XPS analysis reveals that nitrogen doping-induced Ti-C/N bond electron rearrangement strengthens covalent interactions among functional groups, simultaneously improving conductivity and structural stability. This interface engineering strategy offers an innovative solution for flexible electronics development.

Figure 3: Electrical performance testing of MP-TENG.

As shown in Figure 3, the team innovatively developed a self-powered TENG based on the MP nanofiber membrane. The MP-TENG combines porous MP-NM with Ecoflex, achieving contact-force- and frequency-dependent electrical output. Under contact-force drive, output voltage, current, and transferred charge increase linearly with pressure, peaking at 128 V, 8.4 mA m⁻², and 130 μC m⁻², owing to significantly expanded actual contact area. Frequency response tests show output current rising with vibration frequency (2.5 mA m⁻² to 14.25 mA m⁻²), while voltage and charge remain stable due to weak sensitivity of charge transfer rate to motion speed. Load regulation experiments reveal a peak power density of 5.5 W/m² at 2 MΩ external resistance, with efficient capacitor charging capability (2.2 μF/8.8 V or 10 μF/5 V). Durability tests confirm 96% output retention after 10,000 cycles (2 Hz), validating long-term stability. This characteristic offers a self-powering solution for, especially suitable for low-power, intermittent operation scenarios.

Figure 4: Application of MP-NM smart gesture recognition system.

Meanwhile, the research team developed a self-powered gesture recognition system based on MP-NM triboelectric materials. The system utilizes TENG to collect dynamic pressure signals in real-time and achieves high-precision classification through a multilayer perceptron (MLP) neural network. After training on 600 sets of gesture data, the system attained 96.36% recognition accuracy for six international standard gestures, with extendable capability for handwritten character classification (>99% accuracy). The breathability of MP-NM ensures wearing comfort, while its hydrolytic properties enhance biocompatibility, advancing human-computer interaction toward more natural and eco-friendly development. This technology can be further extended to applications including flexible electronic skin, medical rehabilitation monitoring, and industrial human-machine collaboration, driving the evolution of HCI toward imperceptible, naturalized, and intelligent interactions while providing fundamental support for immersive experiences in the metaverse era.

Paper link: https://doi.org/10.1016/j.cej.2025.165703