Photochemically responsive polymer films enable tunable gliding flights

Stimuli-responsive soft materials can morph under external stimuli (heat, humidity, light), enabling multimodal actuation suitable for microrobotics. However, centimeter-scale hovering flight requires high power density and bandwidth, typically necessitating electrically driven piezoelectric/dielectric actuators with onboard power. Light-responsive materials promise wireless control and miniaturization but lack the power and stability for sustained hovering. Nature offers passive wind-dispersal strategies (parachutes like dandelions, gliders like Javan cucumber, and autorotating maple samaras) that achieve stable, efficient dispersal. Prior artificial microfliers inspired by parachutes and gliders demonstrate passive flight but limited in-flight tunability of aerodynamics. This work asks whether light can directly reconfigure wing geometry to dynamically control gliding performance using a single photochemical polymer, enabling contactless modulation of descent speed, rotation, and trajectory without complex electromechanical systems.

The study builds on extensive aerodynamic research of wind-dispersed seeds and bioinspired passive fliers. Dandelion-inspired devices reproduce separated vortex ring dynamics and can achieve light-controlled takeoff/landing, but pappus fragility limits payload capacity. Glider and rotary mechanisms offer robustness and higher loading, driving development of rotary passive fliers and light-driven rotary robots; yet most lack in-flight tunability of aerodynamics. Recent microfliers integrate electronics (e.g., origami parachutes) for mid-flight geometry control, but at the cost of complexity. Photoactive liquid crystalline polymers present an alternative, enabling stand-off, wireless actuation via light with potential for contactless control of passive flight dynamics. The authors also reference comparative load-carrying performance across flight modes and prior demonstrations of photo-actuated soft robots, highlighting a gap in simple, single-material solutions for in-flight aerodynamic tuning.

Materials and photomechanical actuator: Azo-LCN films were prepared by photo-polymerization of a liquid crystal mixture containing 52 mol% RM105, 18 mol% RM23, 21 mol% diacrylate crosslinker RM82, 6 mol% 4,4'-Bis[9-(acryloyloxy)nonyloxy]azobenzene, and 1.5 mol% photoinitiator. A splayed alignment was achieved by assembling a LC cell with one PVA-rubbed substrate (uniaxial) and one polyimide-coated substrate (homeotropic). A 20 μm spacer gap was used. The mixture was infiltrated at 90 °C, cooled to 50 °C (1 °C min⁻¹), and polymerized under 420 nm light (50 mW cm⁻², 30 min). The film exhibits trans–cis isomerization under UV and cis–trans under visible light, with bistable deformation at room temperature.

Artificial maple samara fabrication: The azo-LCN film (homeotropic top, uniaxial bottom) was laser-trimmed to the maple samara wing contour. A natural maple seed root was glued at the base; a polydomain, undeformable LCN strip was attached along the leading edge to tune chordwise mass distribution. Dimensions/masses: wing length 2.5 cm; LCN strip 1.8 cm × 0.26 cm × 0.02 cm; masses: seed root 24 mg, LCN strip 1 mg, LCN wing 7.1 mg; total artificial seed mass 33.1 mg (natural comparator: 34 mg).

Optical characterization and actuation: UV–Vis spectra of solid films showed absorption peak shifts from 360 nm (trans) to 460 nm (cis) upon switching illumination. Trans-isomer population kinetics were measured under varying UV intensities (e.g., 4 to 155 mW cm⁻²). Photochemical deformation kinetics and tip displacement were recorded under multiple UV intensities; deformation occurs near-isothermally. Cis-state lifetime (~300 min) was determined, establishing bistability.

Free-fall and aerodynamic measurements: Natural and artificial seeds were released in still air and crosswind conditions. Terminal descent velocity (Vt), spinning rate (Ω), cone angle β, deflection angle α, and descent factor (Dr = W/(ρSVt²)) were measured. Load-bearing tests added weights at the root to assess Vt–weight trends and aerodynamic efficiency versus wing load (Lw = W/S). UV/visible illumination (360 nm UV; 460 nm visible) was used to preprogram shape before release or to switch mid-air using a visible light zone located ~1 m below a 2 m release point.

Wind tunnel experiments: A vertical wind tunnel (cardboard tube with base fan, Arduino-controlled speed; meshes and honeycomb straightener) generated updrafts. A thin fiber tethered the seed near its mass center to constrain lateral motion. The seed self-stabilized at a height where lift balanced weight; changes in wind speed, UV intensity, and wavelength (UV vs visible) modulated height (h) and Ω. Time-resolved h and Ω were recorded; cyclic reversibility under alternating UV/visible illumination was tested.

Dispersal studies: Indoor crosswind setup used a horizontal fan (1 m height, 0.93 ± 0.07 m s⁻¹). Landing positions (X–Y) and dispersal distances were recorded for natural seeds, artificial seeds pre-UV (visible-relaxed), and artificial seeds post-UV actuation across payloads. Outdoor tests released seeds from 14 m height with ambient wind ~4 ± 0.5 m s⁻¹; landing distances were tracked over 30 trials. Optical reflector payloads (3M tape) facilitated tracking; additional payloads (pH indicator, humidity-responsive polymer) were demonstrated.

Scaling and generalization: Artificial samaras of span 3, 1.5, 0.75, and 0.3 cm were fabricated by laser cutting; Vt and ΔVt after UV were measured. Three additional flight modes were prototyped using azo-LCN: (1) Javan cucumber-like glider with center-of-gravity at mid-wing; localized UV bending on left/right enabled trajectory steering; (2) parachute comprising three azo-LCN strips with transparent triangular membranes; UV-induced closing increased Vt; (3) artificial dandelion with two symmetric pappus assemblies around an azo-LCN hub; UV closed the pappus, while weak white light (50 mW cm⁻²) reopened it to enable takeoff in a 100 cm s⁻¹ updraft.

Data handling: Measurements reported as mean ± standard deviation with n = 3 for repeated-sample tests unless noted; repeatability over multiple light cycles was assessed.

• Bioinspired artificial maple samara achieved autorotation comparable to natural seeds: artificial mass 33.1 mg (vs 34 mg natural), descent speed 98.4 cm s⁻¹ with cone angle β ≈ 15°; natural 72.1 cm s⁻¹, β ≈ 17°.
• Load-bearing tests showed Vt and Ω increased with added weight; descent factor Dr versus wing load was comparable between artificial and natural seeds, indicating similar aerodynamic efficiency across loadings.
• Photochemistry: UV–Vis peak shifted from 360 nm (trans) to 460 nm (cis). Trans fraction decreased to <50% after 120 s at 4 mW cm⁻² UV and to ~50% within 10 s at 155 mW cm⁻². Cis-state lifetime ~300 min at room temperature enabled bistable shape retention without continuous power.
• UV actuation decreased wing deflection angle α (flattening), reducing Vt from 98.4 to 88.2 cm s⁻¹. As UV dose increased, both Vt and Ω initially decreased (higher aerodynamic efficiency, increased Dr), then increased again when α became negative (>3×10⁴ J m⁻² UV dose), mapping a tunable response range.
• Wind tunnel: Under a 1.1 m s⁻¹ updraft, UV irradiation increased stabilized height h and reduced Ω for the photoresponsive seed, with maximal height change Δh increasing with UV intensity; control (polydomain, non-responsive) showed no change. Reversible up–down transitions were achieved by toggling UV (360 nm, 150 mW cm⁻²) and visible light (460 nm, 400 mW cm⁻²). Performance was repeatable over at least 25 light-tuning cycles (gliding) and 100 actuation cycles (strip) without observed decay.
• Mid-air modulation: In free-fall through a visible-light zone, descent speed increased by ~17% relative to pre-UV state, consistent with Vt changes from UV/visible switching.
• Indoor crosswind dispersal (0.93 ± 0.07 m s⁻¹): mean distances—artificial (visible-relaxed) ≈ 15 cm; artificial after UV ≈ 24 cm; natural samara ≈ 30 cm. Landing distributions confirmed light’s impact on dispersal. Increasing payload reduced dispersal distance while maintaining optical tunability.
• Outdoor dispersal (14 m release height; wind ~4 ± 0.5 m s⁻¹): unilluminated artificial seeds landed ~7 m away; UV-illuminated seeds traveled farther, with a mean increase of ~3 m across 30 experiments.
• Scaling: Miniaturized samaras (3, 1.5, 0.75, 0.3 cm) all autorotated with Vt ~100 cm s⁻¹; UV reduced Vt across all sizes, demonstrating scalability of light tunability.
• Generalization: Javan cucumber-like glider exhibited steerable left/right trajectories via localized UV bending. Parachute design showed UV-induced closing that increased Vt, up to ~60% higher than the initial flat state. Artificial dandelion closed under UV and reopened under weak white light (50 mW cm⁻²) to take off in a 100 cm s⁻¹ updraft.
• Actuation speeds during relaxation reached ~100° s⁻¹ (blue-light relaxation), adequate for in-flight modulation during gliding.

The study demonstrates that light fields can act as stand-off, wireless controllers for passive microfliers by morphing aerodynamic surfaces rather than generating lift. Photochemical actuation in azo-LCNs, with near-isothermal operation and long cis-state lifetimes, provides bistable shape states that persist without continuous illumination, minimizing onboard power needs. By tuning the wing deflection angle α, light modulates descent factor Dr, terminal velocity Vt, and spinning rate Ω, enabling reversible control of stabilized height in an updraft and modulation of dispersal distance indoors and outdoors. This directly addresses the core research questions: a single polymer component can reconfigure wing geometry and steer gliding performance dynamically without integrated electromechanical complexity. The approach is robust over repeated cycles and scales from millimeter to centimeter spans, and it generalizes across multiple passive flight modes (autorotation, sail-like gliding, parachuting, pappus-based flight). While photoactuation frequencies are insufficient for hovering flight, they are well suited for adjusting control surfaces during gliding. Future performance improvements could arise from reducing structural mass and optimizing drag–payload trade-offs (e.g., using porous LCNs), and from integrating sensing payloads for environmental monitoring.

The work introduces a light-responsive artificial maple samara that matches natural autorotational behavior and, crucially, enables reversible, bistable, contactless control over flight dynamics (Vt, Ω, altitude) using UV/visible light. Light-mediated shape morphing modulates dispersal indoors and outdoors and remains effective after miniaturization across an order of magnitude in size. The concept extends to other passive flight modes (Javan cucumber-like glider with steerable trajectories, parachute with tunable descent speed, and an artificial dandelion with light-gated takeoff/landing). These results establish a platform for distributed microfliers whose aerodynamics can be tuned remotely, suggesting opportunities for swarm-level control and environmental sensing. Future research should pursue lighter, possibly porous LCN architectures to reduce terminal velocities, optimize response kinetics and durability, and co-design photonic control strategies and payload integration for real-world deployment.

• The approach is not suitable for sustained hovering: required actuation frequencies (>10² Hz) and lift demands exceed the capabilities of current photo-responsive LCN bending/flapping, as discussed by lift scaling and material actuation rates.
• Performance depends on availability and intensity of specific light wavelengths (UV for trans→cis and visible/blue for cis→trans), which may limit operation under variable ambient conditions without auxiliary illumination.
• Demonstrated payload capacity is on the order of tens of milligrams; heavier payloads reduce dispersal distance, indicating a trade-off between functionality and range.
• Indoor and outdoor dispersal tests were conducted under specific wind conditions and geometries; generalizability across diverse atmospheric conditions and terrains requires further validation.
• Although bistable, the cis-state relaxes over hours (~300 min), setting a time window for shape retention without continuous light; long-term material aging and very high cycle counts were not exhaustively characterized.