Weave-pattern-dependent fabric piezoelectric pressure sensors based on polyvinylidene fluoride nanofibers electrospun with 50 nozzles

The study addresses how textile weave structure affects the piezoelectric pressure-sensing performance of fabrics made from PVDF nanofiber yarns. Wearable piezoelectric devices benefit from flexible organic materials like PVDF that can be electrospun into nanofibers with high β-phase content for strong piezoelectricity. Prior fabric-based sensors often relied on melt-spun PVDF requiring post-poling and offered limited dimensional control and sensitivity. This work proposes a sequential, scalable process—50-nozzle electrospinning with on-belt fiber alignment, drawing, and twisting into yarns—followed by weaving with PET warp yarns to form different weft rib patterns (1/1, 2/2, 3/3). The hypothesis is that weave pattern influences contact mechanics and compressive strain distribution at warp–weft crossover points, thereby modulating electromechanical coupling and sensor sensitivity. The purpose is to optimize weave pattern for high sensitivity and demonstrate large-area, wearable, all-fabric pressure sensors capable of detecting diverse human motions. The importance lies in scalable fabrication (to ~195 cm × 50 cm), high sensitivity, and robust wearable applicability without additional poling treatments.

PVDF and its copolymers are widely used in wearable piezoelectric devices due to flexibility and processability, with electrospinning enabling in situ β-phase formation. Nanofiber structures and structural modifications have improved energy harvesting in prior works. Fabric-based approaches using PVDF yarns have been explored: a PVDF yarn fabric energy harvester showed 0.14 V peak voltage and 28 mV N⁻¹ sensitivity after poling (~2.7 V μm⁻¹). A pressure-sensitive fabric woven from PEDOT-coated PVDF yarns achieved up to 18.4 kPa⁻¹ at 100 Pa. Other reports include plain PVDF/Ag-nylon fabrics with 55 mV N⁻¹ sensitivity under 70 N at 1 Hz and knitted PVDF fabrics with ~17.6 mV N⁻¹ at 0.1 MPa. However, many prior textile studies used melt-spinning requiring extra poling and showed less dimensional controllability; continuous electrospinning-based twisted yarns for fabrics have been rarely demonstrated, and high sensitivity without post-treatment has been limited.

PVDF nanofibers were produced using a customized multi-nozzle electrospinner with 50 needle spinnerets (hole ~0.57 mm). A 19 wt% PVDF solution (Kynar flex-2801-00) in a DMSO/acetone mixed solvent was stirred overnight at 60 °C. Electrospinning was conducted at ~20 kV with a feeding rate of 0.2 mL min⁻¹ onto a moving conveyor belt (conveyor feeding rate ~20 m min⁻¹ in schematic). Parallel Cu rods spaced ~1.5 cm apart were mounted across the belt to electrostatically aid fiber alignment; long-range alignment was further induced by pulling toward a rotating drum. The as-spun nanofiber mat (fiber diameters ~700–1300 nm, avg ~973 nm) was merged into a single filament (~120 μm diameter) and then converted into a four-ply yarn (~300 μm diameter) by drawing (draw ratio 2.35), setting, and twisting. Twist levels ranged from 50–500 twists per meter (TPM); mechanical testing identified an optimal tensile strength (~486 MPa) at 300 TPM with rupture strain ~80%. Three woven fabric structures were produced on a rapier loom (weft insertion speed 50 picks min⁻¹, reed width 50 cm): 1/1 (plain), 2/2, and 3/3 weft rib weaves, using PVDF yarn as weft and PET yarn (~100 μm) as warp. Large-area fabrics up to ~195 cm × 50 cm were fabricated. SEM, FFT analysis of alignment (ImageJ), XRD (Cu Kα) with peak deconvolution to estimate β-phase (~55.8%), and FTIR (700–1800 cm⁻¹) characterized morphology and crystallinity. Pressure sensors (2 cm × 2 cm) were built by sandwiching the fabric between an ITO/PET bottom electrode and an Ag-coated nylon fabric top electrode; Cu lead wires were attached and the device encapsulated with PI. Output voltage and current were measured under finger pressing (~1 Hz) using a nanovoltmeter (Keithley 2182A, 10 MΩ) and galvanostat (IviumStat, 1 MΩ), respectively. Applied force (4–24 N typical) was monitored with a FlexiForce A201 tactile sensor attached to the device. All-fabric sensors (4 cm × 4 cm) used Ag-coated nylon for both electrodes and were integrated into garments or insoles; a 12 cm × 12 cm arrayed sensor had screen-printed Ag electrodes (4 × 3 pixels, 1.5 × 1.5 cm² each) sandwiched between cotton cloths. Shoe insole sensors targeted heel and forefoot regions with screen-printed electrodes. Repeatability (up to 1,000 cycles at ~16 N), stretching response (up to ~25% strain), and washability (5 machine washes) were evaluated.

• Weave pattern strongly affected sensor performance. The 2/2 weft rib delivered the highest sensitivity: voltage sensitivity 83 mV N⁻¹ and current sensitivity ~5.0 nA N⁻¹, outperforming 1/1 (24 mV N⁻¹; ~2.8 nA N⁻¹) and 3/3 (36 mV N⁻¹; ~3.4 nA N⁻¹). The 83 mV N⁻¹ is ~245% higher than 1/1 and exceeds typical quartz sensors (10–20 mV N⁻¹).
• Peak outputs increased approximately linearly with force (4–24 N). For the 2/2 fabric, maxima reached ~2.0 V and ~109 nA at 24 N. For 1/1, ~0.62 V and ~63 nA at 24 N; for 4 N, ~0.09 V and ~7 nA.
• Stretching along the PVDF weft yielded a small voltage (~4.5 mV at ~25% strain), indicating limited sensitivity to pure tensile strain.
• The sensors detected diverse physiological/mechanical inputs from ~0.02 N (2.025 g block drop/pick) up to ~694 N (running), with clear signals for bending, twisting, crumpling, finger bending (currents ~15–50 nA for ~30°–120°), walking, and running.
• Large-area, crosstalk-free 4 × 3 pixel arrays showed uniform responses with average ~0.51 ± 0.04 V per pixel upon finger touches.
• Shoe insole implementation resolved gait phases (heel-strike, support, toe-contact, leg-lift) in voltage signals.
• Mechanical and structural properties: four-ply yarn had optimal tensile strength ~486 MPa at 300 TPM; rupture strain ~80%. XRD/FTIR confirmed strong crystallinity with β-phase fraction ~55.8%.
• Durability: after five standard machine washes, the all-fabric sensor retained ~81.3% of its original output current; repeatability maintained over 1,000 pressing cycles.

The superior performance of the 2/2 weft rib fabric is attributed to an optimized balance between the number of warp–weft contact points (higher in 1/1) and the magnitude of compressive strain at each contact (higher in 3/3). This balance enhances electromechanical coupling in the PVDF weft, yielding higher voltage and current sensitivities. The linear force–output relationships facilitate calibration for wearable sensing across varied force magnitudes. Compared to commercial quartz sensors and prior PVDF textile sensors, the achieved sensitivity is high, while the devices remain flexible, breathable, and scalable to meter-scale fabrics via a 50-nozzle electrospinning and standard weaving process. Robust detection of a wide range of inputs (0.02–694 N), crosstalk-free pixelated arrays, and wash durability underscore applicability for self-powered wearable electronics and health monitoring (e.g., gait analysis). The limited response to pure tensile strain emphasizes that compressive interactions at weave crossovers primarily drive signal generation.

This work demonstrates scalable fabrication of PVDF nanofiber yarns via 50-nozzle electrospinning with Cu-rod-assisted alignment, followed by drawing/twisting into four-ply yarns and weaving with PET warp into 1/1, 2/2, and 3/3 weft rib fabrics. The optimized 2/2 pattern achieved high pressure sensitivity (83 mV N⁻¹; ~5.0 nA N⁻¹), linear response up to 24 N, and robust detection of diverse physiological motions from 0.02 N to 694 N. Large-area arrays exhibited uniform, crosstalk-free signals, and devices retained ~81.3% output after five washes, highlighting practicality for wearable, self-powered sensing and gait monitoring. The results indicate that controlling weave architecture can tune contact mechanics and compressive strain to maximize piezoelectric coupling in fabric sensors, enabling production-scale applications.

• Limited tensile strain sensitivity: stretching along the PVDF weft produced only ~4.5 mV at ~25% strain, indicating the sensor is far more responsive to compressive inputs than to pure tensile deformation.
• Washability was assessed for five standard machine-wash cycles with ~81.3% current retention, indicating some performance degradation; longer-term laundering and extended cycling beyond 1,000 compression cycles were not reported.
• Force application during some tests relied on manual finger pressing, which may introduce variability, though a tactile sensor was used to monitor applied force in characterization.