Large-Scale Nanofiber Manufacturing| Self-powered all-nanofiber Janus textile E-skin sensor with air permeabilityand anti-fouling for human-machine interactions

With the development of the Internet of Things, e-skin sensors urgently need to break through limitations in sensing reliability, miniaturization, and portability. The emergence of smart textiles integrated with triboelectric nanogenerators (TENGs) has opened new possibilities for advanced sensors with unparalleled lightweight, breathability, flexibility, and washability.

Recently, Professor Linxi Hou's team at Fuzhou University fabricated an all-nanofiber Janus textile with a multilayer structure using continuous electrospinning and electrospraying as a self-powered e-skin sensor.  After modifying the triboelectric layer with fluorinated polyurethane, the TENG output reached 356 V, 2.88 μA, 80.12 nC, and 2.49 W m−2, capable of powering small electronics via a rectifier circuit.

As a proof-of-concept, a Janus textile-based smart system was developed, achieving precise robotic hand control and material identification using circuit modules and deep learning (1D-CNN), providing tactile recognition for smart robotics and human-machine interaction (HMI). The study, titled "Self-powered all-nanofiber Janus textile E-skin sensor with air permeability and anti-fouling for human–machine interactions," was published in Nano Energy.

Figure 1 illustrates (a) the fabrication process of Janus textiles and (b) TPU/FPU nanofibers. The smart Janus textile applications include (c) TENG, (d) self-powered e-skin sensor, and (e) human-machine interaction.

The Janus textile fabrication process: First, a PLGA nanofiber membrane inner layer was prepared via electrospinning as a substrate for the AgNWs electrode layer, offering good biocompatibility with skin.  Next, the AgNWs electrode layer was deposited on the PLGA nanofibers via electrospraying, forming interconnected conductive pathways.

Simultaneously, an FPU-modified TPU (TPU/FPU) triboelectric layer was electrospun onto the AgNWs electrode. The TPU/FPU nanofiber membrane served as the outer functional layer, generating excellent triboelectric signals and self-cleaning properties.

FPU modification enhanced anti-fouling performance, increased the electronegativity and dielectric properties of the triboelectric layer, and created an excellent tribo-negative material. Based on the Janus textile e-skin's self-powered sensing capability, a robotic hand could accurately mimic human gestures.

Additionally, combined with a 1D-CNN, the smart Janus textile achieved high-precision material identification. The through-pore structure of the all-nanofiber Janus textile also contributed to superior air/water permeability.

Figure 2 presents: (a) SEM image of TPU/FPU nanofibers. (b) XPS survey spectrum of TPU/FPU nanofibers with (c) high-resolution C1s and O1s spectra. (d) Mechanical strength and (e) anti-fouling performance of TPU/FPU nanofibers. (f) Water vapor transmission rate of Janus textile.

Figure 3 (a) Working mechanism of the triboelectric nanogenerator (TENG) based on Janus textile. (b) Potential excitation. (c-e) Open-circuit voltage (Voc), short-circuit current (Isc), and short-circuit transferred charge (Qsc) depending on FPU content. (f) Triboelectric performance of Janus textile-based TENG in contact with eight triboelectric materials. (g) Charge density when contacting different triboelectric materials. (h) FOM DM histogram of eight pairs of triboelectric materials.

Figure 4. Triboelectric performance of the Janus textile-based triboelectric nanogenerator (TENG) depending on variables: (a-c) applied force at 3 mm separation distance and 3Hz contact frequency; (d-f) separation distance at 10N impact force and 3Hz contact frequency; (g-i) frequency at 10N impact force and 3 mm separation distance.

Figure 5. Triboelectric performance of the Janus textile-based TENG depending on: (a) humidity and (b) strain; (c) voltage across the resistor and (d) peak power density curve; (e) schematic of the rectifier circuit; (f) charging curve of commercial capacitors powered by the Janus textile-based TENG and (g) multiple charge-discharge curves; (h) photograph of small electronic devices powered by the Janus textile-based TENG; (i) comparison of triboelectric performance with previous studies; (j) operational stability of the Janus textile-based TENG.

Figure 6. (a) Flow chart of the 1D convolutional neural network (1D CNN) for material classification identification and (b) confusion matrix. (c) Training process of signal classification. (d) Scatter plot of eigenvalue classification. (e) Schematic diagram of the human-machine interaction (HMI) platform. (f) Photos of the HMI process and corresponding output signals from each Janus textile-based triboelectric nanogenerator (TENG) sensor.