Soft, tough, and fast polyacrylate dielectric elastomer for non-magnetic motor

The study addresses limitations of existing dielectric elastomers used in soft actuation, notably high stiffness (Young’s modulus ~0.2–1.0 MPa for commercial acrylics), high driving electric fields (>80 MV m⁻¹), slow response (bandwidth <10 Hz), and high mechanical loss from viscoelasticity. Based on Maxwell stress, actuation strain scales with εr/Y and E², motivating materials with higher dielectric constant and lower modulus. Prior approaches (high-εr fillers, plasticizers, and reduced hysteresis) often incur trade-offs such as increased stiffness, breakdown susceptibility, plasticizer leakage, or reduced εr. The research aims to design a polyacrylate dielectric elastomer that concurrently achieves low modulus, high toughness, low mechanical loss, and satisfactory dielectric properties to reduce driving field and enhance speed, enabling high-performance DEAs and non-magnetic motors.

The paper reviews common strategies to improve actuation: raising dielectric constant via high-εr fillers (which often requires high loading, increasing stiffness and risking breakdown) and lowering modulus via plasticizers (risking leakage and instability). Attempts to reduce hysteresis can diminish dielectric constant. Commercial acrylics like VHB™ 4910 exhibit high dielectric constant (~4.4 @1 kHz) and large strain but demand high driving fields and show creep and high mechanical loss (tan δm ~0.5). Recent work includes molecular grafting, ionic liquid loading in silicones, copolymers, and composite strategies; however, achieving simultaneous low modulus, high εr, and low mechanical loss remains challenging. This study proposes macromolecular crosslinkers and controlled network architectures as an alternative route without fillers or plasticizers.

Design and network optimization: A quantitative structure–property relationship (QSPR) approach was used to select monomer and crosslinkers. n-Butyl acrylate (nBA) was chosen for flexibility and working temperature; polyether-based oligomers were selected as crosslinkers for flexibility and solubility. To create a uniform network and reduce crosslinking point functionality, a difunctional urethane acrylate oligomer CN9021NS (Mn ≈ 28,000 g mol⁻¹; flexible polyether diol + aliphatic diisocyanate) was used, targeting an average molecular weight between crosslinks in the 10⁴–10⁵ g mol⁻¹ range. Comparative networks used polyethylene glycol diacrylate (Mn ≈ 575), CN9893NS (Mn ≈ 1600), and CN9014NS (Mn ≈ 6800) to make BA-S, BA-M, and BA-L. Synthesis (UV curing): Precursors comprising nBA monomer, crosslinker (CN9021NS for BAC series; or small/medium/large crosslinkers for BA-S/BA-M/BA-L), and photoinitiator (2-hydroxy-2-methylpropiophenone) were mixed. An oxygen-free cell (silicone spacer between release films and glass) was filled with precursor, degassed (vacuum, 15 min), UV-cured (2.5 W cm⁻², 3 min), then vacuum dried at 40 °C for 24 h. Film thickness was controlled via spacer. Gel swelling and free-chain analysis: 1-mm-thick 10×10 mm samples were swollen in THF for 1 week to determine swelling ratio and gel fraction; free chains were extracted and analyzed by elemental analysis (PERKIN ELMER CE-440) and GPC (THF; Shimadzu LC20/RID-20) to assess composition and molecular weight distribution. Dielectric characterization: Broadband dielectric spectroscopy (Novocontrol) from 100 Hz to 1 MHz on 1-mm-thick samples; at least three specimens; report εr and tan δ (e.g., εr = 5.75, tan δ = 0.0019 @1 kHz for BAC2). Mechanical testing: ISO 37 standard; 1-mm-thick dumbbell (2 mm × 12 mm gauge) stretched at 200 mm min⁻¹ for stress–strain and Young’s modulus (slope at 5% strain). Cyclic tests on 10 mm × 50 mm strips (gauge 30 mm), stretched to 3× at 100 mm min⁻¹ to obtain mechanical loss (tan δm from loop area). DMA (TA Q800): 1 Hz, <2% strain, −45 to 50 °C at 7 °C min⁻¹. Electrical breakdown: Pillar-to-plate electrodes (25-mm diameter) in silicone oil; DC ramp 500 V s⁻¹ (BDJC-50 kV). Weibull analysis over 10 specimens for characteristic breakdown strength. Actuation tests (static area strain): Films without pre-strain: 1-mm-thick films on annular PMMA frame (inner diameter d = 20 mm) with carbon grease compliant electrodes; voltage increased stepwise (1 kV) until breakdown; hold 1 min per step. Central invagination depth h measured by laser sensor (HL-G105-S-J); area strain S = (2h/d)² × 100%. With 400% equiaxial pre-strain: films on 120-mm frames with 25-mm circular electrodes; field from 0 to 70 MV m⁻¹ in 10 MV m⁻¹ steps, 30 s per step; strain from video via MATLAB script. Dynamic response: With 400% pre-strain at 40 MV m⁻¹ held for 10 min; three repeats; normalized time to 90% final strain noted. Frequency response: 1-mm films on 20-mm frames with single-wall carbon nanotube electrodes; sinusoidal drive 10 Vpp with 5 V offset amplified ×500 (AMT-5B20) to 5 kVpp; laser displacement sensor (LK-g80) measured h; S normalized to 1 at 1 Hz; frequency 1–100 Hz. Cyclic actuation: Same as frequency response at fixed 5 Hz; 50,000 cycles; z-displacement tracked. Non-magnetic motor fabrication and testing: Films equiaxially pre-stretched 4×, mounted on PMMA ring (inside diameter 140 mm) with four electrodes forming quadrants; center ring (inside diameter 40 mm) maintained pre-strain and held gears. Driving commutation via single-chip microcontroller and relays; rotation captured by camera and analyzed frame-by-frame. Torque measured using a string-and-basket over a pulley, increasing weight to loss of synchronization at each field/frequency; power = torque × rotational speed. Transmission gears used for enhanced speed in some tests. Minimum driving electric field and speed–frequency curves recorded for VHB™ 4910 and BAC2 motors.

• Materials performance (BAC2): Young’s modulus ~0.073 MPa (overall 0.073–0.161 MPa across samples); elongation at break ~2400%; toughness 6.77 MJ m⁻³; ultimate strength 32.2 MPa; mechanical loss tan δm = 0.21 @ 1 Hz, 20 °C; dielectric constant εr = 5.75 and dielectric loss tan δ = 0.0019 @ 1 kHz.
• Free-chain contribution: Swelling indicates highest dissociative content (~18%) in BAC2; extracted free chains comprised unreacted CN9021NS and uncrosslinked poly(nBA). Presence of free chains increases εr (BAC2 drops from 5.75 to 5.4 after extracting free chains) without increasing dielectric or mechanical loss; storage modulus and tan δ nearly unchanged after swelling.
• Actuation sensitivity and static actuation: Actuation sensitivity β = εr/Y reaches 78.8 for BAC2 vs 21 for VHB™ 4910. Without pre-strain at 15 MV m⁻¹: BAC2 area strain 18.5% vs 4.5% for VHB™ 4910. With 400% equiaxial pre-strain at 70 MV m⁻¹: BAC2 area strain 118%, ~3.5× VHB™ 4910. Energy density at 70 MV m⁻¹: 0.242 MJ m⁻³ (BAC2) vs 0.042 MJ m⁻³ (VHB™ 4910).
• Dynamic response and stability: At 40 MV m⁻¹ with 400% pre-strain, time to reach 90% of final strain: 35.2 s (BAC2) vs 327.6 s (VHB™ 4910). Frequency response (1–100 Hz, large-signal 5 kV drive): BAC2 nearly flat up to 100 Hz; VHB™ 4910 declines sharply. BAC2 sustained 50,000 cycles at 5 Hz with stable z-displacement.
• Non-magnetic motor performance: Minimum driving electric field reduced from 48 MV m⁻¹ (VHB™ 4910) to 32 MV m⁻¹ (BAC2). At 48 MV m⁻¹, maximum rotation rate 0.72 r s⁻¹ (BAC2) ≈ 15× VHB™ 4910. With transmission gears, BAC2 reached 2.86 r s⁻¹. Output torque and power of BAC2-based motor were ~6× and ~18× those of VHB™ 4910-based motor, respectively.
• Overall: The macromolecular crosslinker strategy yields a soft, tough, low-loss elastomer enabling large strain at lower fields, higher energy density, fast response, and superior motor performance.

By engineering the crosslinked network using a difunctional macromolecular urethane acrylate (CN9021NS) with flexible polyether backbones, the elastomer reduces crosslinking point functionality and enhances network homogeneity. This decreases Young’s modulus and mechanical loss while maintaining or improving dielectric properties. The presence of dissociative (uncrosslinked) chains increases dipole density and enhances orientational polarization under field, boosting εr without adding fillers or compromising losses or modulus. Consequently, the actuation sensitivity β is substantially increased, directly reducing the driving electric field required for a given strain and enabling faster, more stable electromechanical response with reduced creep. The improved electromechanical performance translates to practical gains in soft non-magnetic motors: lower minimum operating fields, higher attainable speeds over broader frequency ranges, and increased torque and power. These results address the key bottlenecks of acrylic DEAs—high drive voltage, slow response, and high viscoelastic loss—offering a materials-centric solution that avoids drawbacks of filler loading or plasticization.

The work introduces a polyacrylate dielectric elastomer with an optimized macromolecular-crosslinked network that simultaneously achieves low modulus, ultra-high toughness, low mechanical loss, and favorable dielectric properties. This combination yields large actuation strain at reduced fields (up to 118% at 70 MV m⁻¹), higher energy density, fast and stable response up to and beyond 100 Hz, and superior soft motor performance (lower threshold field, up to 15× speed, and significantly higher torque/power versus VHB™ 4910). The strategy—matching crosslinker molecular dimensions to network mesh and leveraging dissociative chains to enhance polarization—provides a generalizable path to high-performance DEAs without fillers or plasticizers. Potential future research directions include: mapping resonance behavior beyond 100 Hz and optimizing device geometries for high-frequency operation; improving drive electronics (higher current HV amplifiers) to fully exploit material speed; exploring other monomers and macromolecular crosslinkers to tune β and loss; scaling fabrication and long-term reliability testing under diverse environments; and integrating these elastomers into complex soft robotic and biomedical actuator systems.

• Measurement bandwidth: Frequency response was measured up to 100 Hz; resonance onset at 100 Hz suggests peaks may lie beyond the tested range.
• Drive electronics constraint: For BAC2 motors, output was limited by the experimental platform’s high-voltage amplifier (charging time constant ~10⁻² s), not solely by material properties; higher-power amplifiers could further improve performance.
• Pre-strain requirement: Maximum strains and motor demonstrations relied on ~400% equiaxial pre-strain, which may complicate some applications.
• Free-chain quantification: While beneficial effects of dissociative chains are shown, precise long-term stability and potential migration/leaching under varied conditions were not fully explored.
• Breakdown statistics: Weibull methodology described but specific characteristic breakdown strengths are not reported in the main text.