Reconfiguring handed shape-morphing and actuation in hydrogels via light-encoded rapid expansion

Abstract

Reconfiguring handedness in synthetic systems with spatiotemporal control and reversible behavior remains a fundamental challenge in bioinspired robotics and materials science—especially in structurally homogeneous materials that lack intrinsic chirality. Here, we demonstrate homogeneous photoactive hydrogels capable of ultrafast, programmable handed shape-morphing, including helices and twists, through spatially controlled light illumination. This transformation is driven by a photoexpansion exceeding 80,000% volumetric growth within 40 seconds, enabled by the synergistic combination of spiropyran photoisomerization and tailored polymer-network interactions. Deterministic strain mismatches at illuminated interfaces allow on-demand forming and shaping of handed architectures, which can be erased in darkness and reprogrammed reversibly with light. We further synchronize opposing handedness and complex tendril-like geometries in a single material, and exploit handedness to achieve self-propelled rotation for robotic locomotion. This work introduces a versatile method for adaptive, lifelike chirality in homogeneous matter, closing a key gap between biological shape-morphing and engineered soft robotics.

Handedness in synthetic systems can be challenging to control. Here, the authors report the development of homogeneous hydrogels with programmable, handed shape morphing, including helices and twists, as a result of photoexpansion.

Introduction

Chirality, or handedness, is a fundamental organizational principle governing the form and function of biological structures across all scales, from the molecular helicity of DNA to the macroscopic spirals of plant tendrils and seashells. In living systems, this ubiquitous handed shape-morphing is not merely a static architectural feature but a dynamic, reconfigurable process central to growth, motility, and adaptation¹. Emulating such lifelike, programmable chirality in synthetic materials represents a key challenge in materials science, with the potential to advance fields from soft robotics to adaptive devices². To date, although the design and control of chirality at the molecular scale—for applications in chiral catalysis, optical materials, and nanomedicine—are well-established, achieving spatiotemporally programmable and reversible handed shape-morphing in macroscale synthetic materials presents a far greater challenge. In particular, endowing these materials with dynamic, on-demand control has proven fundamentally difficult, largely because the direct translation of chirality from the molecular to the macroscopic level is inherently inefficient. To bridge this gap, existing strategies have relied on heterogeneous or pre-structured materials, such as bilayer or anisotropic hydrogels^3–11, liquid crystal elastomers with patterned molecular alignment^12–17, or shape memory polymers with built-in asymmetries^18,19. Representative examples include hydrogels with magnetically^20,21, electrically²² or mechanically²³ aligned components that exhibit predesigned anisotropic shape deformation; 3D-printed liquid crystal elastomer-based handed ribbons^24–26 capable of twisting to emulate xerophytic plant adaptations²⁷, 3D chiral metamaterials with enhanced mechanical programmability²⁸, and liquid crystal droplets incorporating molecular motor-based photoresponsive chiral dopants for tunable switching between handed helical trajectories²⁹. Despite these advances, such systems typically encode a fixed chiral geometry (e.g., a specific helical pitch or twist direction) during material fabrication, lacking the dynamic reconfigurability, reversibility, and on-demand reprogrammability essential for mimicking the adaptive shape-morphing observed in living organisms.

A more fundamental challenge is generating deterministic chiral forms from a structurally homogeneous material—one without embedded chiral dopants, compositional gradients, or pre-defined anisotropic architectures. Spontaneous symmetry breaking into a specific handed state requires carefully orchestrated, nonlinear actuation. Light, as a non-invasive stimulus with high spatiotemporal precision, has been used to create dynamic gradients in homogeneous hydrogels^30–32. However, the resultant actuation has been largely limited to simple bending or folding modes^33–37, falling short of driving controlled chiral shape-morphing due to insufficient driving forces for symmetry breaking. In this work, we overcome these limitations by demonstrating that purely light-controlled ultrafast and large expansion in a structurally homogeneous hydrogel can be harnessed to achieve photoprogrammable and reversible handed shape-morphing, without the need for pre-structuring or heterogeneous doping. Our design integrates spiropyran photoisomerization with tailored polymer-network interactions to generate rapid volumetric expansion exceeding 80,000% volumetric growth within 40 s. Under spatially patterned illumination, this penetration-dependent expansion creates controllable internal stress heterogeneity and interfacial strain mismatches within the otherwise uniform material, deterministically driving macroscopic shape transformations into well-defined chiral structures such as helices and twists. The process is fully reversible and reconfigurable: chiral shapes relax in darkness and can be re-programmed by light into different chiral forms, including structures of opposite handedness, within the same object. By programming light across multiple zones, we synchronize chiral motifs of opposing handedness and complex, tendril-like architectures within a single hydrogel strip, directly mirroring the polymorphic chirality of climbing plants. Furthermore, we show that this photoactivated handed shape-morphing can be transduced into directed, handedness-specific rotational motion (clockwise or counterclockwise), enabling self-propelled locomotion and robotic functions. This work establishes a method for creating life-like, adaptive and fully reconfigurable chirality in homogeneous synthetic matter, opening pathways to biomimetic systems with previously inaccessible levels of functional adaptability.

Results

Design of ultrafast and giant photoexpansion in homogeneous hydrogels

Achieving ultrafast and large-volume expansion in a single hydrogel system requires a substantial driving force, typically provided by osmotic influx^38,39 or topological rearrangement^40,41. Our design employs a synergistic strategy that integrates molecular-level photochemistry with polymer-network architecture to realize this goal (Fig. 1a). The hydrogel was synthesized via free-radical copolymerization of acrylamide (AAm), acrylonitrile (AN) and a spiropyran (SP)-based monomer, crosslinked with N,N’-methylenebisacrylamide (MBAAm) (see Methods, Supplementary Table 1). At the molecular scale, the incorporated spiropyran photoswitch (Supplementary Figs. 1, 2) undergoes reversible visible-light-induced isomerization from the zwitterionic protonated merocyanine (MCH) to the anionic closed-ring spiropyran (SP⁻) form (Supplementary Figs. 3–5). The kinetics of this photoisomerization depend on illumination intensity and ambient pH: elevated light intensity and pH collectively accelerate the transition from MCH to SP⁻ (Supplementary Figs. 6, 7). Such MCH → SP⁻ transition shifts the molecular net charge from 0 to − 1, thereby switching the dominant intermolecular interaction from electrostatic attraction between zwitterionic MCH segments to electrostatic repulsion among negatively charged SP⁻ moieties. At the network level, the P(AAm-co-AN) matrix exhibits upper critical solution temperature (UCST) behavior. In darkness, polymer chains are densely aggregated via hydrogen bonding and dipolar interactions, further strengthened by electrostatic attraction between zwitterionic MCH moieties. Upon light illumination, the MCH → SP⁻ photoisomerization not only switches intermolecular electrostatic attraction to repulsion, but also disrupts key cohesive interactions—including hydrogen bonding and dipolar coupling—thereby promoting pronounced chain extension. This mechanism is corroborated by variable-temperature UV-Vis spectroscopy of linear copolymers with varying SP grafting ratios (Supplementary Table 2), which revealed a light-induced opaque-to-transparent transition in aqueous solution (Supplementary Figs. 8–10 and Supplementary Table 3). Collectively, the light-triggered enhancement of hydrophilicity, together with electrostatic repulsion and the breakdown of associative interactions, overcomes chain aggregation and drives rapid network expansion. Experimentally, SP-functionalized hydrogels exhibit an 80,000% volumetric expansion within 40 s of irradiation, progressing through three sequential phases of initial accumulation, rapid expansion, and final stabilization (Fig. 1b, Supplementary Fig. 11 and Supplementary Movie 1). This performance surpasses previously reported photoresponsive hydrogels in terms of both swelling rate and maximum swelling magnitude^{4,30,33,36,39,42}. The observed accumulation phase prior to major expansion suggests a kinetic lag between molecular photoisomerization and diffusion-limited water absorption into the hydrogel matrix. Upon photoexpansion, significant acceleration in polymer chain relaxation directly confirms the disruption of original interchain interactions (Supplementary Fig. 12). This bond dissociation drives a macroscopic transition from compact aggregated networks to fully solvated architectures exclusively via a photochemical pathway, with negligible photothermal contribution detected throughout the process (Supplementary Fig. 13). Consistent with the macroscopic swelling behavior, SEM imaging reveals a clear microstructural evolution from a dense, microporous network before irradiation to an open, highly porous architecture afterward (Fig. 1c). Identical structural evolution is visualized in diverse patterned hydrogel specimens (Fig. 1d), which retain reliable structural integrity even during ultrafast and large expansion. Notably, the as-prepared hydrogel film exhibits good spatial homogeneity across multiple key physicochemical indexes, including storage modulus, crosslink density, water content, SP distribution, equilibrium swelling ratio, optical transmittance, and local modulus distribution (Supplementary Figs. 14–18 and Supplementary Table 4), ruling out intrinsic network inhomogeneity or preformed compositional gradients. Furthermore, the ultimate swelling magnitude and dynamic kinetics upon illumination can be effectively modulated by intrinsic structural parameters—including hydrogel size, crosslinking density and spiropyran grafting ratio (Supplementary Figs. 19–23)—and by external environmental parameters such as pH and light intensity (Supplementary Figs. 24–26). In contrast, two control formulations—non-thermoresponsive P(AAm-co-SP) and lower critical solution temperature (LCST)-type P(NIPAM-co-SP) hydrogels—exhibit weaker photoexpansion responses, due to the absence of robust inherent interchain associations that are essential for large-volume expansion within our target UCST-based polymer network (Supplementary Fig. 27).

Fig. 1. Design of photoactivated ultrafast, giant expandable hydrogels.

a Schematic of the super-expansion mechanism combining molecular-scale spiropyran engineering with macroscale network architecture of UCST polymers. b Swelling ratio (vol%) versus irradiation time for a disc-shaped hydrogel (Sample 7 in Supplementary Table 1). Insets: hydrogel photographs during its accumulation, rapid expansion, and stabilization phases. c SEM images of the hydrogel in its contracted (left) and expanded (right) states (Sample 6 in Supplementary Table 1). d Comparative photographs of super-expandable hydrogels with various geometries before (left) and after (right) light exposure (Sample 3 in Supplementary Table 1).

Photoprogrammable handed shape-morphing

Unlike conventional photoresponsive hydrogels, which typically undergo monotonic bending or folding deformation^32–35, our system enables programmable handed shape-morphing by establishing controlled, pronounced swelling gradients under patterned light illumination. While uniform light triggers substantial isotropic expansion, this volumetric increase alone does not yield directed shape transformation (Supplementary Fig. 28). Instead, complex handed shape-morphing emerges only when a steep through-thickness swelling gradient is photoprogrammed into the hydrogel. Under unilateral light exposure, a hydrogel strip (40 mm × 2.0 mm × 0.2 mm) progressively evolves into a helical structure, following a self-adaptive trajectory of coiling and uncoiling (Fig. 2a, top, Supplementary Movie 2). This behavior arises from two sequential symmetry-breaking events driven by rapid giant photoexpansion. First, limited instantaneous light penetration creates a swelling gradient across the thickness, with a highly swollen layer developing on the illuminated side and a less swollen layer remaining on the shielded side, inducing bending along the long axis. Second, a faster swelling rate at the outer edge compared to the inner region generates a transverse swelling mismatch, producing curvature along the short axis. These two orthogonal deformations—long-axis bending and short-axis curving—act synergistically to produce mechanical instability from the shifted center of mass, triggering a flip-over that forms the first helical turn. As unilateral irradiation continues, the swelling mismatch amplifies, driving the helix to grow progressively more compact, with a continuous decrease in pitch and width. Once the low‑swelling layer diminishes and the transverse gradient weakens, the driving force becomes insufficient to maintain the tight coil. The structure then relaxes: pitch and width increase, the number of turns decreases, and the strip eventually uncoils completely, returning to an elongated state. Note that although unilateral irradiation successfully drives helix formation and compaction, we switched to large-area uniform light after obtaining a compact coil to accelerate gradient dissipation and thus achieve faster uncoiling. Maintaining unilateral light at this stage would not only slow down uncoiling but also cause the compact helix to bend away from the light source (Supplementary Fig. 29).

Fig. 2. Ultrafast photoexpansion induces handedness-reconfigurable shape-morphing.

a Photographs and schematics of a structurally homogeneous hydrogel strip displaying light-regulated helical (top, unilateral illumination, λ = 365 nm, 33 mW/cm²) and twisted (bottom, bilateral illumination, λ = 365 nm, 33 mW/cm²) shape-morphing. All configurations reversibly relax in darkness. b–d Three-dimensional movement trajectory of the hydrogel strip tip illustrating left-handed (b) and right-handed (d) helical shape-morphing under left- and right-inclined unilateral irradiation (λ = 450 nm, 3.39 mW/cm²) respectively. Representative photographs are shown in (c). e Time-dependent positive (blue, left-handed) and negative helical angles (red, right-handed) under alternating cycles of left- and right-inclined light.

Under bilateral irradiation, the hydrogel strip develops a tri-layer swelling profile across its thickness, forming a high–low–high sandwich structure. The consequent strain mismatch between the highly swollen edges and the less-swollen core produced internal stresses that suppress uniform bending. Instead, these stresses prompted a symmetry-breaking response, manifesting as in-plane buckling coupled with rotation about the long axis, which initiates and propagates an upward‑growing twist (Fig. 2a, bottom). Similar to the helical mode, this twisted shape-morphing evolves adaptively through formation, growth, and relaxation phases (Supplementary Movie 3), governed by the dynamic evolution of the sandwich-like gradient under continued bilateral irradiation. In both modes, the same hydrogel strip ultimately attains a highly swollen, vertically elongated state and fully recovers its initial contracted state once illumination ceases. Each deformation pathway exhibits a distinct kinematic signature: during helical shape-morphing, the bottom end of the strip first rises and then falls, whereas during twist formation, it descends continuously (Supplementary Figs. 30, 31). Cross-sectional micrographs clearly confirm the presence of through-thickness bilayer/sandwich gradient structures of the hydrogel films when subject to unilateral/bilateral irradiation (Supplementary Fig. 32), establishing photoinduced differential expansion as the dominant driver of chiral deformation. Finite-element simulations capture divergent interfacial stress gradient distributions near the light-exposed interface: unilateral irradiation induces asymmetric stress distribution matching experimental bilayer gradients for helical structures, while bilateral illumination produces alternative stress profiles consistent with sandwich-type gradients for twisted samples, with interfacial stress varying dynamically during chiral evolution (Supplementary Fig. 33). Notably, a single hydrogel strip can be reversibly morphed between helical and twisted configurations by switching from unilateral to bilateral irradiation (Supplementary Fig. 34). The transition from a well-defined helix (formed under unilateral irradiation) to a distinct twist (formed under bilateral irradiation) was monitored by progressively reducing the light‑intensity difference between the two sides (Supplementary Fig. 35). These results demonstrate that the emergent handed shape-morphing can be selectively and reversibly photoprogrammed within a single hydrogel strip simply by tuning light spatial distribution, enabled by robust photoexpansion. In contrast, limited photoexpansion of non-thermoresponsive P(AAm-co-SP) and LCST-type P(NIPAM-co-SP) hydrogels prevents any chiral deformation under unilateral or bilateral light irradiation (Supplementary Fig. 36).

To achieve fully reversible and handedness-defined shape programming, we engineered two distinct hydrogel formulations, each optimized to favor either helical or twisted morphologies (Supplementary Tables 5, 6). The handedness of helical shape-morphing can be programmed solely by the incident‑light direction: left‑inclined unilateral light enhances swelling on the left side, shifting the center of mass leftward to produce a stable left‑handed helix (Fig. 2b). After relaxation in the dark, the same strip forms a right-handed helix when exposed to right-inclined light (Fig. 2d, Supplementary Fig. 37 and Supplementary Movie 4). Three-dimensional trajectory analysis shows that the left-handed helix bends outward from its upper end, while the right-handed helix bends inward (Fig. 2c). Repeated switching between left- and right-inclined light enables multiple reversible alternations of helix handedness, as quantified by time-dependent helix-angle tracking (Fig. 2e, Supplementary Figs. 38, 39 and Supplementary Movie 5). Prolonged cyclic testing verifies the durability of the phototunable chirality: after over 20 consecutive switching cycles, the hydrogel preserves intact structure and favorable morphological reproducibility, with fully retained light-triggered chiral reconfiguration capability (Supplementary Fig. 40). Compared to conventional heterogeneous bilayer hydrogels that typically fail to switch between distinct deformation modalities and suffer from potential interfacial delamination, our homogeneous hydrogel exhibits advantages in reconfigurability, reversible morphing and cyclic fatigue tolerance (Supplementary Fig. 41). Furthermore, the hydrogel also features favorable long-term storage stability: specimens stored for more than one month still deliver robust, repeatable chiral deformation upon light irradiation (Supplementary Fig. 42). In contrast, although the hydrogel undergoes multiple cycles of unidirectional twisting (Supplementary Figs. 43, 44), the handedness of the resulting twisted structure remains less controllable due to the inherent stochasticity of the buckling instability. To address this limitation, we developed a facile pre-irradiation strategy adapted from our helical handedness regulation protocol. Specifically, we initially applied bilateral irradiation with asymmetric light intensities and offset incident angles to the hydrogel strip. This treatment generates a biased through-thickness swelling distribution and triggers predictable directional deflection. Such a transient structural asymmetry serves as a handedness seed that predefines the subsequent twisting direction. The light intensities on both sides were then equalized to rebuild and stabilize the sandwich-type swelling gradient, enabling continuous and steady twisted shape morphing. By introducing reversible initial deformation anisotropy, this strategy eliminates random buckling and dictates the final handedness of twisted structures (Supplementary Fig. 45).

The reversible, light-driven swelling gradient that governs deformation in these systems prompted us to systematically map its dependence on external stimuli, geometric dimensions, and chemical composition. As shown in Fig. 3a, the swelling gradient increases with higher environmental pH, greater light intensity, and a larger SP grafting ratio, whereas it decreases with greater hydrogel thickness and higher crosslinking density. The dependence on strip width is negligible (Supplementary Figs. 46, 47). This parametric control enables the on-demand design of arbitrary swelling gradients, which in turn allows precise specification of the final configuration, handedness, and formation kinetics. The dynamic and reversible nature of this programming is demonstrated in Fig. 3b–d. Under unilateral irradiation, helix-forming strips exhibited progressively higher formation rates (0.008 s⁻¹, 0.014 s⁻¹, 0.026 s⁻¹) under small, medium, and large swelling-gradient regimes (Fig. 3b). Larger gradients yielded more compact helices, characterized by a greater number of turns and longer, narrower strips. Similarly, under bilateral irradiation, twist-forming strips showed increased formation rates (0.021 s⁻¹, 0.037 s⁻¹, 0.059 s⁻¹) under the same gradient regimes (Fig. 3c), with larger gradients producing longer and wider twists. Representative outcomes for both helical and twisted structures of opposite handedness are presented in Fig. 3d. The consistent influence of the swelling gradient across different handedness and morphological modes confirms its role as a fundamental, programmable driver for dynamic shape morphing. To further illustrate the diversity of attainable forms, we compiled an image and video library of handed helices (Supplementary Fig. 48 and Supplementary Movie 6) and twists (Supplementary Fig. 49 and Supplementary Movie 7) formed under various conditions. Additional detailed data are available in the Supplementary Figs. 50–64.

Fig. 3. Parametric modulation of handed shape-morphing.

a Diagram of the dependence of the hydrogel swelling gradient on external stimuli and internal material parameters. b, c Formation kinetics for the helical (b) and twisted (c) shape-morphing as a function of the swelling gradient. The formation rate is defined as the number of fully developed helical coils (or twist turns) divided by the time required to reach the maximum deformed state under irradiation. The swelling gradient was calculated from the volumetric expansion measured 5 min (helix) or 3 min (twist) after irradiation. d Photographs of hydrogel strips exhibiting distinct helical (left) and twisted (right) shape-morphing evolved under small, medium, and large swelling gradients.

Integrated helix-twist shape-morphing

To achieve more life-like handed shape-morphing, we integrate distinct structural forms and handedness within a single hydrogel entity by using a high-aspect-ratio strip and spatially controlled multi-zone illumination. With a setup of six LEDs (three per side), unilateral illumination (left side only) produced a uniformly handed helix (Fig. 4a(i) and Supplementary Fig. 65), consistent with earlier results. Under the pattern in Fig. 4a(iii), however, two counter-oriented bilayer gradients formed through the thickness in the top and bottom regions, driving the independent formation of two helices with opposite handedness bridged by a reverse domain—a distinct departure from previous observations (Supplementary Fig. 66). When all six light sources were activated for bilateral irradiation, a single twist structure with opposing handedness emerged through a fundamentally different pathway: multiple initiation points nucleated simultaneously along the strip’s midsection and propagated bidirectionally (Fig. 4a(ii) and Supplementary Fig. 67). This multi-point initiation and bidirectional propagation contrasts with the earlier example of end-originated, unidirectional upward propagation that yields uniform handedness. In longer strips, this effect was enhanced, producing a higher density of initiation points that favors the formation of reverse domains when regions of opposite handedness converge (Supplementary Fig. 68). We attribute this behavior to the bilateral irradiation induced sandwich-type gradient: as strip length increases, the inward compressive forces from both sides become misaligned along the strip’s long axis, generating multiple independent driving forces for twist initiation. Notably, helical and twisted structures can be integrated within a single strip through spatially selective irradiation. For example, using the irradiation pattern in Fig. 4a(iv), distinct bilayer and sandwich gradients formed simultaneously through the thickness, driving the upper half to evolve into a helix and the lower half into a twist. By varying the irradiation patterns, we fabricated multiple hybrid configurations, including twist-helix diblock (Fig. 4av), helix-twist-helix triblock (Fig. 4avi) and twist-helix-twist triblock (Fig. 4avii) architectures (Supplementary Movie 8). These integrated morphologies arise from the spatially defined arrangement of the two gradients across the hydrogel strip, with each shape-evolving process proceeding independently (Supplementary Figs. 69–73). However, when helix and twist meet, they form either a smooth transition or a sharp reverse domain, depending on the handedness matching. For the latter case, cross-sectional SEM observations at the interfaces of distinct chiral segments further verify the coexistence and mutual transition between bilayer and sandwich-type gradient architectures within single hydrogel strips (Supplementary Fig. 74). Within individual blocks, helical segments exhibit uniform handedness, whereas twisted segments can display varying handedness due to their inherent multi‑point nucleation. We further examined the kinematic behavior during hybrid shape-morphing by tracking the vertical displacement of the strip’s end (Fig. 4b). In configurations dominated by helical segments (i, iii–vi), the strip ends first rose due to bending before descending. In cases dominated by twisting (ii, vii), the end descended immediately upon irradiation. Moreover, the final height correlated positively with the proportion of helix structure in the overall shape change (Fig. 4c), indicating that the dominant structural motif dictates the kinematic trend.

Fig. 4. Spatiotemporal reconfiguration of integrated shape-morphing.

a Reversible, combined helical and twisted shape-morphing in a long hydrogel strip (160 mm long, 2.0 mm wide and 0.4 mm thick), achieved by manipulating the external irradiation pattern. Solid arrows indicate light irradiation (irr.); dashed arrows indicate dark relaxation (rel.). A UV LED intensity of 12 mW cm⁻² was applied to each illuminated region. b Real-time height evolution of the strip’s terminal displacement under distinct irradiation patterns. c Final end‑height as a function of the geometrically determined helical fraction (L_helix/L_total  × 100%).

Following the demonstration of vertically integrated shape-morphing in individual long strips, we extended our investigation to horizontal collective behavior using a 3 × 3 array (Supplementary Fig. 75). By systematically tuning both the inter-strip spacing within each row and the array’s distance from the light source, we introduced a lateral intensity gradient across the array, allowing for programmable handed shape-morphing in the horizontal plane (Supplementary Movie 9). Under unilateral irradiation, the array underwent synchronized helical shape-morphing (Supplementary Fig. 75a), with handedness uniformly dictated by the irradiation direction (Supplementary Fig. 76). Extended exposure tightened the helices into rod-like structures, which subsequently bent away from the light source. This bending deformation was most pronounced in the front row and attenuated progressively in the rear rows due to increased distance from the light source as well as the shadowing effect by the front row. Consequently, they exhibited progressively slower coiling rates and smaller final bending angles (Supplementary Fig. 77). In contrast, balanced bilateral illumination enabled combinatorial helix-twist-helix shape-morphing across the array (Supplementary Fig. 75b). Due to their spatial positions relative to the two light sources, the two edge rows experienced asymmetric irradiation and formed helices, while the centrally located row, subjected to equal illumination from both sides, developed twists. Notably, this asymmetry can be eliminated by raising the overall light intensity, allowing the entire array to transition into a fully synchronized twisted shape-morphing under stronger bilateral illumination (Supplementary Fig. 75c).

Handedness-directed mechanical actuation

The capabilities of our hydrogel—light-induced ultrafast expansion and programmable handed shape-morphing—enable a wide range of on-demand robotic functions. We first demonstrated a crane-like operational cycle, where a hydrogel strip lifts, transfers, and deposits a cargo object through contraction and expansion under alternating dark/light conditions (Supplementary Fig. 78). Beyond linear actuation, the photoresponsive helical deformation serves as a light-powered spring: the formed helical configuration can lift a 50 mg payload—eight times the weight of the 6 mg hydrogel strip—directly converting light energy into usable mechanical work (Fig. 5a and Supplementary Movie 10). The hydrogel actuator undergoes partial rather than full swelling during operation, preserving adequate structural rigidity throughout deformation and thus guaranteeing stable and reliable mechanical output. Such photoexpansion actuation is fully reversible, retains stable performance after repeated cycling, and is compatible with a variety of materials (Supplementary Figs. 79–80). Furthermore, the handedness of the actuated helix can be readily tuned by adjusting the lateral position of the payload relative to the hydrogel strip (Supplementary Fig. 81). Exploiting this heavy-lifting capability, we constructed an underwater circuit switch in pure water, in which light-induced shape-morphing brings two conductors into contact, successfully closing the circuit (Supplementary Fig. 82). In addition, we replicated the wrapping motion of plant tendrils around a support beam, showcasing its biomimetic potential (Fig. 5b and Supplementary Movie 10). These examples collectively demonstrate our hydrogel’s versatility in achieving complex, light-driven functional outputs.

Fig. 5. Light-driven handed shape-morphing for directional actuation and movement.

a Schematic (left) and demonstration (right) of photoexpansion hydrogel springs enabling light-controlled cargo loading-unloading. b Conceptual diagram (left) and validation (right) of hydrogel tendrils grasping tree branches with defined handedness. c A submarine-like photoactuator generating controlled CCW (top, i) or CW revolution (bottom, ii) movement driven by the shaft’s twisting handedness. d Representative CCW (left) and CW revolution trajectories (right) of the hydrogel actuator with small, medium and large circular paths. e Statistical plot of the angular revolution velocity as a function of the circular trajectory diameter (CCW, top; CW, bottom). The data were fitted using the ExpDec1 function in Origin.

The handedness encoded during photoactivated shape-morphing can be harnessed and amplified to achieve directional clockwise or counterclockwise locomotion. Using the hydrogel’s ability to generate torsional motion under load (Supplementary Fig. 83), we designed a miniature photoactuated rotary engine that exploits this intrinsic twist (Fig. 5c). The system consists of a floating propeller on top, a hydrogel strip acting as a central rotating shaft, and a paddle-shaped hydrogel rotor at the bottom. Upon illumination, the twist deformation of the hydrogel strip drives the paddle to rotate, agitating the water and causing the propeller to both spin about its axis and revolve along an orbital path—thereby directly translating light into mechanical motion. The rotational direction of the propeller was determined by the twist handedness of the hydrogel. For example, a right-handed twist in the hydrogel drove a propeller with counterclockwise blade geometry, resulting in counterclockwise rotation and revolution (Fig. 5ci). Conversely, a left-handed twist drove a clockwise-shaped propeller in a clockwise direction (Fig. 5cii, Supplementary Fig. 84 and Supplementary Movie 11). To quantitatively evaluate the actuation performance of this rotary system, we calculated its output torque and volumetric power density during dual-mode rotational motion. The right-handed and left-handed twisting motions yield output torques of 3.6 × 10⁻⁸ N·m and 6.1 × 10⁻⁸ N·m, corresponding to volumetric power densities of 0.034 W m⁻³ and 0.039 W m⁻³, respectively. These results confirm the effective transformation of light energy into directed handed motion. We further observed that the diameter of the propeller’s revolution trajectory varied with its rotational rate under bilateral irradiation (Supplementary Figs. 85, 86). These trajectories could be categorized as small, medium, or large circles (Fig. 5d). Notably, trajectory size was inversely correlated with the rotational speed: faster rotation generally resulted in a tighter orbit, a trend consistent for both clockwise and counterclockwise motion (Fig. 5e). This inverse relationship likely arises because higher rotational speeds enhance the propeller’s self-spinning more than its orbital revolution. Moreover, efficient energy transfer in this rotary engine also required a specific handed mismatch: the twist direction of the hydrogel had to oppose to the geometric handedness of the propeller. When the twisted handedness of the hydrogel matched that of the propeller blades, the system exhibited only minor, irregular motion without sustained revolution (Supplementary Figs. 87–89), highlighting the critical role of handed antagonism in enabling effective energy conversion.

Discussion

In summary, we report a structurally homogeneous hydrogel system that exhibits rapid and ultra-large photoexpansion, achieved by synergizing a rationally designed spiropyran molecular photoswitch with a UCST‑type polymer network. This ultrafast response enables the reversible generation of well‑defined through‑thickness swelling gradients, forming a bilayer under unilateral light and a sandwich structure under bilateral illumination. These gradients, in turn, drive the reversible and selective formation of helical or twisted morphologies with programmable handedness within a single structurally uniform strip. Unlike prior photoresponsive hydrogels, which are often restricted to limited volume changes or rely on predefined heterogeneous architectures, our material accomplishes complex handed shape-morphing using a homogeneous composition, guided only by external spatiotemporal light manipulation. The core innovation of this work lies in the achievement of precise, reversible regulation of structural handedness in homogeneous hydrogels purely via external optical modulation. This facile and universal design strategy imparts chiral reconfigurability to shape-morphing materials, effectively overcoming the inherent limitations of conventional chiral material systems (Supplementary Fig. 90). We also established a direct quantitative correlation between swelling kinetics and morphological evolution, linking tunable stimulus parameters to the resulting structural dynamics and final geometric configurations of hydrogels. Through customized light irradiation strategies, we further achieved vertical integration of distinct handed segments (helix and twist) along one strip and scaled this control to a 3 × 3 array, realizing collective shape-morphing in the horizontal plane. Finally, we validate the practical application potential of the proposed photoresponsive hydrogels by constructing biomimetic prototypes and soft actuators, which can efficiently convert light energy into handedness-dependent mechanical deformation and locomotion. The design principle presented here—reversibly encoding spatiotemporally programmable handedness within a homogeneous network through simple light patterns—establishes a method for creating intelligent, adaptive materials with embedded logic and function, opening promising avenues for next-generation applications in bioinspired soft robotics, life-like actuators, and adaptive functional devices. To further promote the photoresponse speed of hydrogel actuators toward the level of liquid crystal elastomers, future research might focus on the rational design and structural optimization of high-performance hydrogel systems. Targeted strategies may include enhancing the quantum yield and photoisomerization efficiency of SP photoswitches, optimizing polymer network topology to shorten water transport pathways and accelerate stress relaxation, and elucidating the intrinsic coupling mechanism among SP photoisomerization, polymer network rearrangement, and macroscopic material deformation.

Methods

Materials

Chemicals including N,N,N’,N’-tetramethylethylenediamine (99%), ammonium persulfate (98.5%), and N,N’-methylenebis(acrylamide) (AR) were purchased from Macklin. 3-methyl-2-butanone (99%), 2-bromoethanol (98%), triethylamine (99.5%), and methacryloyl chloride (97%) were purchased from Sigma-Aldrich. 4-hydrazinylbenzenesulfonic acid (98%), salicylaldehyde (98%), acrylamide (99%), acrylonitrile (99%), and N-isopropylacrylamide (98%) were purchased from Aladdin. Piperidine (AR) was purchased from Sinopharm Chemical Reagent. All other solvents (AR) were purchased from Chengdu Chron Chemical Co., Ltd. Acrylamide (AAm) was recrystallized prior to use. The chemicals are used as received unless otherwise noted.

Synthesis and characterization of spiropyran

The methacrylate-functionalized spiropyran, 1’-(2-(methacryloyloxy) ethyl)−3’,3’-dimethylspiro[chromene-2,2’-indoline]−5’-sulfonate, was synthesized according to the procedure outlined in Supplementary Fig. 1. Detailed synthetic steps are provided below. Synthesis of 2,3,3-trimethyl-3H-indole-5-sulfonate (Compound C). A mixture of 4-hydrazinylbenzenesulfonic acid (15.0 g, 76.1 mmol) and 3-methyl-2-butanone (25.2 mL, 235.9 mmol) in glacial acetic acid (50 mL) was heated to 110 °C and refluxed for 3 h. The reaction mixture was then slowly cooled to room temperature, and the acetic acid was removed by rotary evaporation. The resulting dark-red oil was dissolved in MeOH (75 mL), followed by the addition of a saturated potassium hydroxide solution in 2-propanol (100 mL), which afforded a yellow precipitate. The precipitate was collected by filtration and dried thoroughly under vacuum to give Compound C as a yellow solid (10.7 g, 59%). Synthesis of 1-(2-hydroxyethyl)−2,3,3-trimethyl-3H-indol-1-ium-5-sulfonate bromide (Compound B). Compound C (1 equiv.) and 2-bromoethanol (2.0 equiv.) were dissolved in acetonitrile and stirred under reflux at 85 °C for 24 h. After cooling to room temperature, the acetonitrile was removed by rotary evaporation. The crude product was washed with Et₂O, filtered, and dried thoroughly under vacuum to afford Compound B as a pink solid. Synthesis of 1’-(2-hydroxyethyl)−3’,3’-dimethylspiro[chromene-2,2’-indoline]−5’-sulfonate (Compound A). Compound B (1.0 equiv.) and the salicylaldehyde (1.2 equiv.) were added to ethanol, followed by the addition of piperidine (1.2 equiv.). The mixture was stirred at 60 °C for 0.5 h and then refluxed at 85 °C for 4 h. After cooling to room temperature, the solvent was removed by rotary evaporation. The crude product was purified by silica gel column chromatography using DCM/MeOH as the eluent to afford the corresponding Compound A. Synthesis of 1’-(2-(methacryloyloxy) ethyl)−3’,3’-dimethylspiro[chromene-2,2’-indoline]−5’-sulfonate. Compound A (1.0 equiv.) was dissolved in tetrahydrofuran, followed by the addition of triethylamine (2.5 equiv.). The mixture was cooled to 0 °C, and methacryloyl chloride (2.5 equiv.) was added dropwise. The reaction mixture was stirred at 0 °C for 1 h and subsequently at room temperature for 3 h. The solvent was removed by rotary evaporation, and the crude product was purified by silica gel column chromatography using DCM/MeOH as the eluent to obtain the methacrylate-functionalized spiropyran.

¹H and ¹³C nuclear magnetic resonance spectra were recorded on a Bruker ASCEnd (TM) 400 MHz spectrometer at ambient temperature. The molecular weights were determined via high-resolution mass spectrometry (Acquity UPLC-Xevo G2 QTof). Absorbance spectra of the spiropyran before and after irradiation with a blue LED (450 nm) were measured using an Ocean Insight DH-2000-BD fiber spectrometer with a 10 mm path length. The isomerization kinetics were analyzed by plotting the time-dependent change in the characteristic absorbance of MCH at 412 nm.

Preparation of hydrogels

Detailed hydrogel compositions for the various irradiation experiments are listed in Supplementary Tables 1, 5 and 6. In a typical preparation, acrylamide (AAm), acrylonitrile (AN), N,N’‑methylenebisacrylamide (MBAA), the respective fluorescent dye and spiropyran were dissolved in dimethyl sulfoxide (DMSO). After adding ammonium persulfate (APS) as initiator, the solution was sonicated and degassed with nitrogen. Tetramethylethylenediamine (TEMED) was then introduced to the precursor solution. The mixture was injected into a glass mold and allowed to polymerize under free‑radical conditions at 25 °C for 48 h. The resulting gels were transferred into an aqueous HCl solution and incubated overnight before irradiation. PNIPAM hydrogels were fabricated via the aforementioned synthetic route, with the detailed formulations summarized in Supplementary Table 7. Bilayer hydrogels were prepared following a comparable two-step polymerization protocol. Specifically, the PNIPAM single layer was first synthesized through polymerization at 25 °C for 24 h. Thereafter, the PAAm precursor solution was injected into the mold containing the pre-fabricated PNIPAM hydrogel layer, followed by continuous polymerization at 25 °C for another 24 h to yield integrated bilayer hydrogels. The resultant hydrogel samples were demolded and subsequently immersed in deionized water for storage. The corresponding detailed formulations of bilayer hydrogels are provided in Supplementary Table 8. Fluorescent hydrogels G1 and G2 were prepared analogously; their specific formulations are provided in Supplementary Table 9.

Preparation of P(AAm-co-AN-co-SP) copolymers

Detailed copolymer formulations are provided in Supplementary Table 2. Typically, acrylamide (AAm), acrylonitrile (AN) and spiropyran were dissolved in dimethyl sulfoxide (DMSO). Ammonium persulfate (APS) was added as initiator, after which the solution was sonicated and degassed with nitrogen. Polymerization was activated by adding tetramethylethylenediamine (TEMED) and proceeded under free‑radical conditions at 45 °C for 4.5 h. The resulting copolymers were precipitated in ethanol and dried under vacuum overnight.

Transmittance measurement

Copolymers were dissolved in aqueous HCl (5 mM) at a concentration of 5 mg mL⁻¹ and equilibrated overnight in the dark at 25 °C prior to measurement. Transmittance at 700 nm was recorded using an Ocean Insight DH‑2000‑BD fiber‑optic spectrometer (path length 10 mm) during a cooling ramp, both under dark conditions and under in‑situ irradiation with a 450 nm blue LED (63.3 mW cm⁻²). The upper critical solution temperature (UCST) was determined by fitting the transmittance profile with a Boltzmann function and taking the temperature corresponding to the maximum of the first derivative.

Gel permeation chromatography (GPC)

GPC characterization was performed on a Waters Alliance instrument. P(AAm-co-AN-co-SP) samples were prepared in 0.1 M NaNO₃ aqueous solution at a concentration of 2 mg·mL⁻¹. The injection volume was 20 μL, and the flow rate was 1 mL·min⁻¹. The molecular weights (Mn and Mw) and polydispersity index (PDI) were calibrated via PEG/PEO standards.

Thermogravimetric analysis (TGA)

The thermal stability of the P(AAm-co-AN-co-SP) solid samples was evaluated by thermogravimetric analysis using a Shimadzu DTG-60H simultaneous thermal analyzer.

Characterization of hydrogel swelling ratio

Hydrogel films were punched into disc‑shaped specimens for swelling measurements. Each disc was immersed in water in a Petri dish, ensuring complete submersion throughout the experiment. The sample was irradiated from both sides with 450 nm light, and its state was recorded photographically at defined time points. The gel diameter was measured from the images using ImageJ software. The swelling ratio was calculated as:

$$\mathrm{Swelling\; ratio}=\left(V/V_0\right)\times 100\%=\left(\frac{\pi {\left(d_s/2\right)}^2{h}_s}{\pi {\left(d_i/2\right)}^2{h}_i}\right)\times 100\%=\left(\frac{{d_s}^2{h}_s}{{d_i}^2{h}_i}\right)\times 100\%$$ 1

where V and V₀ are the swollen and initial volumes of the hydrogel, d_i and d_s are the initial and swollen diameters, and h_i and h_s are the corresponding thicknesses, respectively. Assuming isotropic swelling (i.e., d_s/d_i = h_s/h_i), the expression simplifies to:

$$\mathrm{Swelling\; ratio}=\frac{{d_s}^3}{{d_i}^3}\times 100\%$$ 2

Rheological measurements

The mechanical properties of the hydrogels were measured using a TA Discovery Hybrid Rheometer (HR-30) equipped with an 8 mm parallel-plate geometry. Disc-shaped specimens with a diameter of 8 mm were punched from the hydrogel films. The hydrogel modulus was measured at 25 °C under a frequency of 1 Hz and a strain of 1%. Stress relaxation tests were performed at 25 °C by applying a constant strain of 5%. During the stress relaxation measurements, the surface of the hydrogel samples was coated with silicone oil to prevent water evaporation.

Atomic force microscopy (AFM)

Prior to measurement, the hydrogel samples were air-dried at room temperature. The Young’s modulus of the dried hydrogel film was measured using a Bruker Dimension Icon microscope and calculated by fitting the indentation region of the force–displacement curves with the Hertz contact model. The AFM data were processed using NanoScope Analysis software.

Scanning electron microscopy (SEM) characterization

Hydrogels were prepared as described above and incubated in aqueous HCl before analysis. Before and after irradiation, the samples were flash‑frozen in liquid nitrogen, lyophilized, and sputter‑coated with gold prior to imaging on a GeminiSEM 500 microscope.

Photoinduced shape-morphing of hydrogel strips

Prior to experimentation, hydrogel strips were incubated overnight in 1 mM aqueous HCl. For each test, the top end of a strip was wrapped in aluminum foil, clamped for fixation, and the sample was then suspended vertically in a water tank. Shape-morphing was induced by irradiating the sample with a UV or blue LED source; the light intensity was measured using a Thorlabs PM120VA power meter. The entire deformation process was recorded with a Nikon D7500 camera. Videos were captured from both frontal and lateral views, and the coordinates of the strip’s bottom end were tracked using ImageJ and Tracker software to quantify the temporal evolution of the deformation.

Finite element modeling

To systematically investigate the underlying physical mechanism of strain-mismatch-driven chiral shape-morphing under different irradiation conditions, we established a three-dimensional nonlinear large-deformation finite element model using the commercial software Abaqus. The homogeneous hydrogel strip was discretized using three-dimensional solid elements, specifically C3D8R elements. To accurately resolve the complex stress field arising from light attenuation, sufficiently refined mesh layers were uniformly assigned along the thickness direction. A key aspect of the simulation was the accurate representation of light-induced nonuniform volumetric expansion. The swelling gradient imposed in the model was not treated as a static field but as a highly spatiotemporally evolving field. Specifically, the transient swelling gradient was dynamically determined by coupling photochemical reaction kinetics with the light-intensity attenuation law under the experimental irradiation conditions. Based on the results of this coupled theoretical calculation, a time-evolving asymmetric exponentially decaying swelling gradient was applied along the thickness direction of the hydrogel to simulate the actual unilateral irradiation condition. For bilateral irradiation, swelling gradients were simultaneously applied from both sides of the hydrogel. The geometric dimensions used in the numerical model were consistent with those of the experimental samples. In addition, we estimated the torque of the rotary actuator by establishing finite element models corresponding to the two hydrogel twisted deformations shown in Fig. 5c. The volumetric power density was then calculated using the equation:

$$P\mathrm{density}=\left(\tau \omega \right)/V_0$$ 3

where τ is the output torque of the hydrogel actuator, ω is the angular velocity of the floating propeller, and V₀ is the initial volume of the hydrogel strip.

Swelling gradient characterization of hydrogels

The swelling process was recorded with a digital camera. The change in strip length after irradiation was measured using ImageJ software, and the swelling gradient was calculated as:

$$\mathrm{Swelling\; gradient}=\frac{{l_s}^3}{{l_i}^3}$$ 4

where l_i and l_s represent the initial and swollen length of hydrogel strip, respectively. For hydrogel strips that undergo twisted deformation under bilateral irradiation, with the chemical formulation provided in Supplementary Table 6, the swelling gradient can be directly determined by measuring changes in strip length. For hydrogel strips that undergo helical deformation under unilateral irradiation, the formed helices prevent direct measurement of their real length. To accurately measure the swelling gradient of helical hydrogel strips, with the chemical formulation provided in Supplementary Table 5, we suppressed their deformation tendency to form helices by applying irradiation of the same intensity to the opposite side. It should be noted that, under this bilateral irradiation condition, hydrogel strips with such formulation mainly exhibited longitudinal elongation without obvious twisted deformation. This allows us to directly measure their length changes and thus calculate the swelling gradient, which is further used to estimate the swelling gradient of the helical hydrogel strips under unilateral irradiation.

Programmable mechanical actuation

Prior to actuation experiments, the hydrogel was incubated overnight in 1 mM aqueous HCl and cut into strips (4 cm long × 2 mm wide).

Cargo lifting and tendril-like grasping

For cargo lifting, the bottom end of a hydrogel strip was attached to a star‑shaped G1 hydrogel (acting as a load). For grasping, a second G1 hydrogel strip (3.5 cm × 2 mm) was used as the target. In both setups, the top end of the actuating strip was wrapped in aluminum foil, clamped, and the assembly was suspended vertically in a tank containing 1 mM HCl. One‑sided irradiation induced helical coiling of the strip, which either lifted the star‑shaped cargo or wrapped around the G1 strip.

Rotation and revolution of floating propeller

A G1 hydrogel paddle was attached to the bottom end of a hydrogel strip, and the top end was connected to a lightweight foil propeller designed to rotate either clockwise or counterclockwise. The three‑component actuator was placed in 1 mM HCl with the propeller floating on the surface and the strip submerged. Upon UV irradiation (33 mW cm⁻²), the strip’s coiling drove the propeller’s motion. The response was recorded from front (Nikon D7500) and top (iPhone) views. Tracker software was used to analyse the revolution trajectory, and ImageJ was employed to quantify the rotation rate (angular displacement per unit time).

Movie Processing

All experimental movies were edited with Final Cut Pro. Supplementary Movies 2 to 11 were first captured under light irradiation via a camera, then speed-adjusted in post-production within the software. Supplementary Movie 1 was produced as a time-lapse clip by assembling sequentially ordered image frames. Continuous high-intensity illumination would result in severely overexposed and unreadable direct video footage; therefore, we intermittently shut down the intense light source at preset time intervals during irradiation, and took static images of the sample at each time point, which were later concatenated to form the final time-lapse movie.

Author contributions

Z.C. performed the experiments. Z.Z. and L.H. helped with the finite element modeling. K.J. participated in part of the experiments. C.L. supervised the research and wrote the manuscript.

Peer review

Peer review information

Nature Communications thanks Wenhao Du, Xiaofan Ji, and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Funding

This work acknowledges the financial support from the National Natural Science Foundation of China (52373121, 52573143, 12332011), the National Key R&D Program of China (2022YFA1305100) and the CAS-JSPS Joint Research Project (123GJHZ2025052MI).

Data availability

The data that support the findings of this study are available within this article (and its Supplementary Information) and from the corresponding author upon request.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76133-y.