Low-hysteresis, pressure-insensitive, and transparent capacitive strain sensor for human activity monitoring

The paper addresses the need for wearable, stretchable strain sensors that combine low hysteresis, fast response, transparency, and long-term reliability for applications such as human motion detection, soft robotics, electronic skin, and human–machine interfaces. Existing sensors often prioritize stretchability and gauge factor but neglect robustness metrics like hysteresis, stability of electrodes, and decoupling from pressure. Transparency is also important for unobtrusive wear and integration with photosensitive medical components. The authors propose a transparent capacitive strain sensor employing interdigital three-dimensional electrodes made from vertically aligned carbon nanotubes embedded in PDMS to achieve high sensitivity with low hysteresis, pressure insensitivity, and good transparency for accurate human-activity monitoring.

The authors review materials used for stretchable substrates, including PDMS for flexibility and transmittance, Ecoflex and hydrogels for high stretchability, and other elastomers such as epoxy aliphatic acrylate, aliphatic urethane diacrylate, PVDF, TPU, natural rubber latex, SEBS, and polyimide. For electrodes, carbon materials (carbon black, graphene nanoplatelets, carbon nanotubes, graphene) are favored for conductivity and processability; other options include silver nanoparticles, liquid metal, gold, and Ag nanowire/MXene composites. They summarize sensing mechanisms: optical sensors require additional photodetection and depend on transmittance changes; piezoelectric sensors can be limited in range; resistive sensors are simple but suffer from nonlinearity, significant hysteresis, and pressure–strain coupling due to conductivity changes and cracking; capacitive sensors rely on geometric changes of dielectric and electrodes, offering better linearity and low hysteresis. However, conventional capacitive sensors typically use vertical sandwich electrodes that block light and suffer from electrode instability (cracking) during stretching, motivating the interdigital in-plane, embedded 3D electrode approach.

Device fabrication: A 2 nm Fe seed layer was deposited and patterned on a Si wafer with a 1 µm SiO2 layer via e-beam evaporation and lift-off to define interdigital fingers. Vertically aligned carbon nanotube (VACNT) forests were grown by microwave plasma-enhanced CVD to form electrodes. A degassed PDMS precursor (10:1 monomer:curing agent) was spin-coated at 150 rpm for 40 s, followed by vacuum treatment for 20 min to remove bubbles and promote PDMS infiltration into VACNT gaps, embedding the CNTs to form 3D electrodes. PDMS was cured at 70 °C for 2 h, and the PDMS with embedded electrodes (thickness ~400 µm) was peeled off and reattached to the wafer on its opposite side. A top PDMS layer was then spin-coated at 2000 rpm for 40 s and cured at 70 °C for 2 h to encapsulate and protect electrodes; masking tape was used to keep contact pads exposed. The final device was peeled off. Electrode geometries used interdigital patterns with equal width and gap of 20, 50, and 80 µm; 3D electrode height was ~35 µm measured by laser confocal microscopy. Characterization included optical transmittance (380–900 nm) of pure PDMS and PDMS with CNT electrodes; electrical stability of electrode resistance under strain up to 50%; and mechanical–electrical performance under tensile strain using a stretching test setup. Theoretical modeling: A capacitance model for interdigital electrodes was developed, partitioning total capacitance into parallel electrode interactions and end/backbone coupling, and incorporating elastic deformation with PDMS Poisson ratio (v = 0.5). The model predicts normalized sensitivity as a function of strain and geometry (example dimensions l0 = 50 µm, d0 = 100 µm, h0 = 100 µm, w0 = 20 µm). Experimental data were fitted with a semiempirical factor (C_test/C_model = 0.57) to account for fringe fields, parasitics, material properties, and electrode deformation. Performance testing: Sensitivity (gauge factor), hysteresis across full strain range, transient response time via stretch-and-release (to 90% recovery), short-term stability with step inputs, long-term cycling stability (1500 cycles), and pressure coupling were assessed. Pressure-insensitivity was measured from 0 Pa to 1 MPa. Human activity demos included sensor placement on finger, knee, elbow, wrist, and neck, and detection of small strain from mouth opening.

- Transparency: Pure PDMS transmittance at 550 nm ~79%. PDMS+CNT interdigital devices at 550 nm: 61.6% (80 µm gap/width), 57.6% (50 µm), 52.7% (20 µm). Larger gaps yield higher transparency. - Electrode robustness: 3D CNT electrode resistance variation <1.6% over 0–50% strain, indicating no cracking and stable electrodes. - Gauge factors: 0.637 (20 µm), 0.505 (50 µm), 0.413 (80 µm). Smaller gaps yield higher sensitivity. - Hysteresis: Ultrallow maximum hysteresis across full range: 0.35% (20 µm), 0.26% (50 µm), 0.13% (80 µm). - Response time: ~60 ms for transient recovery across tested strains (20%, 50%, 90% strain changes), similar for all gap designs. - Long-term stability: Stable operation over 1500 stretching cycles with no significant performance attenuation. - Pressure insensitivity: Capacitance variation <0.8% under normal pressures from 0 Pa to 1 MPa, indicating low cross talk to normal force. - Electrode geometry: 3D electrode height ~35 µm; PDMS substrate thickness ~400 µm. - Modeling: Theoretical model with Poisson ratio 0.5 matches experiments after applying a semiempirical fitting factor (C_test/C_model = 0.57), valid over 0–100% normalized strain. - Application demos: Accurate monitoring of large motions (finger, knee, elbow, wrist, neck bending) and small motion (mouth opening).

Embedding vertically aligned CNT interdigital electrodes within PDMS creates stable 3D conductive paths that maintain electrode integrity during stretching, minimizing resistance changes and cracking; this underlies the observed ultralow hysteresis and long-term stability. The in-plane interdigital capacitive configuration reduces thickness and improves transparency relative to vertical sandwich designs while also reducing coupling to normal pressure, as evidenced by <0.8% capacitance change up to 1 MPa. The design allows a tunable trade-off between sensitivity and transparency by adjusting electrode gap/width: smaller gaps increase capacitance change (higher gauge factor) but decrease transmittance. The theoretical model that accounts for elastic deformation and geometry captures the device behavior over a wide strain range, supporting design optimization. Collectively, these characteristics directly address the need for robust, low-hysteresis, pressure-insensitive, and transparent strain sensing suitable for precise human motion monitoring and wearable interfaces.

The work demonstrates a transparent, flexible capacitive strain sensor using embedded 3D interdigital VACNT electrodes in PDMS, achieving ultralow hysteresis, pressure insensitivity, fast response (~60 ms), stable electrodes (<1.6% resistance variation up to 50% strain), and durable cycling performance (1500 cycles). Transparency and sensitivity can be balanced via electrode geometry, enabling unobtrusive wearability while maintaining adequate gauge factors. The sensor accurately monitors diverse human activities from large joint motions to subtle mouth opening, indicating strong potential for human motion detection, soft robotics, and medical care. Future research directions are not specified in the provided text.

Explicit limitations are not detailed by the authors. A noted design trade-off exists between transparency and sensitivity as electrode gap/width increases. Pressure decoupling was characterized up to 1 MPa, and long-term cycling was demonstrated for 1500 cycles; broader ranges or extended lifetimes beyond these conditions are not reported in the provided text.