Twist imparts nonlinear shear stress and induces entropy changes in materials, forming the fundamental basis of twistocaloric cooling. This refrigerant-free and energy-efficient approach offers a promising alternative to conventional vapor-compression cooling systems. It provides distinct advantages, including low driving stress, considerable cooling power, and compatibility with flexible devices. Despite increasing efforts, a universal structure-performance relationship linking molecular design to device-level cooling behavior remains elusive. In this review, we summarize recent advances in twistocaloric cooling materials and devices, highlighting the growing demand for twist-based structures in thermal management. We focus on the molecular design and structural characteristics of representative twistocaloric materials, including shape memory alloys, entropy-elastic polymers, and semi-crystalline fibers, and compare their cooling capacities arising from twist-induced stress field and hierarchical molecular alignment. Strategies for enhancing cooling amplitude and coefficient of performance are also discussed alongside emerging device architectures that leverage nonlinear stress fields for high integration, low actuation force, and long cycle life. Further, we present the remaining challenges and further perspectives of the twistocaloric field, emphasizing pathways toward intelligent, mechanically robust, energy-efficient, and environmentally adaptable thermal management technologies.

High-efficiency refrigeration materials with long cycle lifetimes are essential for reducing energy consumption in conventional cooling systems. Twistocaloric cooling, which harnesses nonlinear torsional stress, offers a promising pathway to enhanced cooling efficiency. However, a general design strategy for polymer-based twistocaloric materials that combine high efficiency and long cycle life remains elusive. Here, we report spider silk-inspired polybiurea elastomer fibers that exhibit exceptional mechanical properties and twistocaloric cooling performance. The material's architecture features nanoknot-like domains formed via multiple hydrogen bonds and π-π interactions in the hard segments, which knot together the soft segments to provide high mechanical stability and substantial entropy changes. The polybiurea fibers with nanoknot-like domains achieved an extraordinary combination of breaking strength of 316.5 MPa and toughness of 523.4 MJ m-3. This engineered deformable knotted structure enables a maximum cooling temperature drop of -17.1 K (by twisting and stretching), a maximum Carnot efficiency of 89.7%, and operational durability up to 120 000 mechanical fatigue life cycles. We further demonstrate two out-of-phase operated twistocaloric devices designed to recover input mechanical energy, thereby enhancing overall system efficiency. This work presents a robust materials strategy for advancing high-performance polymeric refrigeration materials and systems.

Natural spider silk achieves its exceptional performance through the synergy of rigid β-crystallite anchors and stretchable, energy-dissipating β-sheet domains. However, perfectly replicating this architecture in biomimetic systems remains a significant challenge. Here, we report the development of super-tough artificial silk fibers utilizing a hybrid covalently anchored network (CGN) architecture based on bioinspired glue-like nanoparticles (GNPs). By integrating these GNP cross-links into a stable covalent network, we created a system where adsorbed polymer chains on the GNPs form dynamic loops with substantial hidden lengths. Upon stretching, these loops unfold and dissipate massive energy via the sacrificial rupture of interfacial bonds, mimicking natural β-sheet domains. Crucially, adjusting GNP cross-link size and interfacial binding strength enables precise tuning of the fiber's mechanical balance. Furthermore, the stable covalent skeleton serves as a topological anchor, imposing entropic penalties on chain desorption. This constraint suppresses polymer detachment from the GNP surfaces, compelling chains to undergo stress-induced orientation and assemble into highly aligned nanofibrils. The resulting fibers achieve a tensile strength of 1.25 GPa and a toughness of 306 MJ m-3, successfully reproducing the strength-toughness synergy of native silk. This approach offers a scalable pathway for engineering high-performance fibers for applications in artificial muscles and advanced actuators.

Electrothermal polymer artificial muscles are increasingly popular in soft robotics due to their high load‐to‐weight ratio and cost‐effectiveness. However, current control strategies for these actuators mostly rely on traditional proportional integral derivative (PID) control or single‐dimensional closed‐loop control. These methods can mitigate the adverse effects of the inherent hysteresis, nonlinearity, and temperature–force angle coupling of these actuators somewhat, but fail to effectively meet the precision requirements of specialized applications. To address these challenges, a temperature self‐sensing‐driven hierarchical closed‐loop fractional‐order proportional integral derivative (FOPID) control scheme is designed. It leverages nickel wire's resistance‐temperature characteristics for real‐time temperature self‐sensing, eliminating the need for external sensors; its hierarchical structure prioritizes temperature control followed by angle regulation to mitigate coupling effects; and the FOPID controller outperforms traditional PID with smaller overshoot and faster convergence to manage nonlinearity. Based on this control scheme, an artificial muscle is fabricated using polyethylene and silver‐plated nylon (PE@SPN). To enhance deformation and output force, single PE@SPN fibers are combined in series–parallel to form composite artificial muscles. A rehabilitation manipulator is developed based on these composite artificial muscles, with a closed‐loop control algorithm implemented via temperature self‐sensing and FOPID as the core control strategy, demonstrating the engineering application value of PE@SPN artificial muscles.

High-strength and tough fibers are essential in biological systems, such as ligaments and tendons, where they preserve tissue integrity under mechanical stress. Spider silk uniquely combines exceptional strength and toughness, arising through the hierarchical self-assembly of spidroin, while also offering excellent biocompatibility and inherent biodegradability, making it an ideal candidate for biomedical applications. Here, we demonstrate that rearranging spidroin peptide chains into ultrafine nanofibrils yields engineered natural spider silk fibers with markedly enhanced both mechanical and functional performance, achieving a breaking strength of 2.0 GPa, toughness of 480 MJ m-3, and actuation stress of 92.6 MPa. Importantly, the contractile actuation enables these fibers to function as a self-adapting surgical suture: beyond passive wound closure, the fibers actively generate postoperative stress, improving tissue approximation and promoting repair. This work establishes a generalizable strategy for engineering natural fibers with integrated mechanical robustness, actuation, and biomedical utility, opening new avenues for intelligent biomaterials design.

Currently, the development of artificial muscles that simultaneously possess high sensitivity, high linearity, and self-sensing capabilities remains a significant challenge. Inspired by the spider’s slit organ, a novel carbon nanotube/liquid crystal elastomer (CNTs/LCE) artificial muscle has been developed. This structure integrates a crack-based sensing unit, a helical deformation mechanism, and self-sensing functionality. A monolithic architecture featuring a helical crack sensor was constructed, which maintains high sensitivity while achieving a large deformation range. In this configuration, the helical structure serves to "kill two birds with one stone": it acts as a sensor that significantly enhances the strain capability, while also functioning as a twisted helical artificial muscle. Furthermore, the introduced crack structure markedly improves sensing sensitivity. When combined with a porous structure that enhances deformability, and utilizing the helical geometry to further amplify the deformation amplitude (up to 110

Hydrogel fibers are promising candidates for artificial tissues, yet achieving high mechanical robustness with high water content remains challenging. Here, we report a deformation-driven strategy that converts strain localization into a controllable pathway for necking created in a nano-assembly in single-component hydrogel fibers. By integrating wet spinning and confinement-assisted drying to form an oriented densified core-sheath structure, followed by programmed necking and salting-out treatment, super strong and tough poly(vinyl alcohol) hydrogel fibers are fabricated. This structure enables spatially confined necking to generate amplified local strain, thereby accelerating chain disentanglement and orientation and promoting nanofibrillar assembly and refined nanocrystalline organization. The resulting fibers exhibit exceptional mechanical performance, reaching a maximum tensile strength of 122.5 MPa, a maximum toughness of 350.4 MJ m-3, and a high water content of 67.9%, comparing favorably with previously reported hydrogel fibers. Moreover, the fibers maintain their shape, sustained mechanical performance, and long cyclic life under simulated body-fluid conditions for 7 days, demonstrating promising potential for artificial ligament application. This study provides a scalable route to high-performance hydrogel fibers and offers new insights into harnessing deformation-enabled structural programming for soft yet robust tissue-mimetic materials.

Inspired by the β-sheet nanocrystals in natural spider silk, we develop a high-damping polycrystalline-phase liquid crystal elastomer (LCE) fiber enabled by a semi-interpenetrating network. Continuous large-scale fabrication of this crosslinked system is realized using a unique channel-confinement strategy. By innovatively designing the end-group molecular structures of linear polymers, we precisely regulate the liquid-crystal phases within the semi-interpenetrating network fibers. Four distinct liquid-crystal phases are constructed, mimicking the β-sheet nanocrystals of spider silk to enable efficient energy dissipation. The resulting fibers exhibit a high elastic modulus of 47.6 MPa, outstanding toughness of 60.4 MJ m-3, a high dissipation coefficient of 88.6%, an ultra-broad damping temperature window, a wide damping frequency range, and a strong actuation stress. When woven into damping nets for impact buffering, the nets exhibit a tunable memory recovery time and an exceptionally low dynamic rebound ratio of 5.9%, enabling efficient impact-energy adsorption and secure capture. Overall, this work overcomes the long-standing trade-off among mechanical, actuation performance, and damping capacity of LCEs, and provides a universal strategy for elastomer-based damper design and precise liquid crystal phase control, opening new opportunities for applications in elastomer dampers, artificial muscles, and soft robotic systems.

Artificial muscle fibers generate heat when connected to direct current, and undergo deformation and mechanical force under the action of thermal effects. In the laboratory, artificial muscle fibers are generally placed perpendicular to the desktop, with weights suspended at the bottom. The amount of mechanical force generated by the artificial muscle fibers is calculated by measuring the deformation after being electrified and the weight of the suspended weights. Traditional pilots use compensatory anti load equipment to achieve compensatory anti load effects by wrapping it around the body surface, generating axial forces directed towards the interior of the body, and squeezing blood vessels/organs. The direction of the force is perpendicular to the direction of the artificial muscle fiber force in the laboratory. Conduct research on the driving methods and actuation modes of artificial muscle fibers, complete the direction conversion of mechanical force of artificial muscle fibers represented by nickel titanium alloy fibers, construct artificial muscle devices, verify the electric thermal driving effect, and form various actuation modes such as compensatory anti load, providing a basis for establishing a new electric driven surface compression scheme.

Self-oscillating actuators that can achieve autonomous motions are highly desired in autonomous soft robotics and intelligent devices. Moreover, oscillators driven by multi-stimuli have attracted considerable interest and have potential applications in multiple complex environmental systems. However, most actuation systems require manual control of switches, and film-based twisting/untwisting oscillation and length stretching/contraction oscillation have not been realized. Here, we fabricated a helical nanofiber composite film and achieved twisting/untwisting oscillation, bending oscillation, and elongation/contraction oscillation under heat, light, and moisture stimuli. Moreover, the oscillator can realize continuous mechanical work under different loads as well as continuous electrical output. This study not only provides the twisting oscillation and twisting motion mechanism but also presents a versatile strategy to fabricate hydrogel-based bilayer hierarchical porous nanofiber composite film twisting oscillators. This actuation system with a twisting self-oscillation load capacity and a multi-stimuli response will be used for autonomous smart devices, autonomous energy conversion, and multi-scenario applications.

Developing high-performance bio-based fibres is highly desirable for improving the sustainability of materials. Cellulose is one of the most abundant bio-derived feedstocks to fabricate such materials. However, the fabrication of high-strength macro cellulose fibres is challenging due to the difficulty in obtaining ordered packing of cellulose molecular chains and nanocrystals in the macro-fibres. Here we develop a draw spinning/de-acetylation method to prepare cellulose fibres with highly ordered molecular packing that incorporates high strength in the obtained fibres. Specifically, a fibre draw spun from well-dispersed cellulose triacetate solution was de-acetylated to generate cellulose fibres, which were then twisted to spirally align the molecular chains. The resulting fibres exhibited mechanical strength of 3.08 GPa and toughness of 215.1 MJ m-3, much higher than existing fibre materials. This work paves the way to obtaining high-performance bio-based fibres.

The rapid development of aerospace, artificial intelligence, and flexible wearable electronics has led to an increasing demand for multifunctional electromagnetic interference (EMI) shielding materials, especially for lightweight and high-strength biomimetic intelligent actuators. In this study, we present polyolefin elastomer/aramid nanofiber/carbon nanotube (POE/ANF/CNT) composites with a sandwich architecture fabricated via layer-by-layer technology. Actuation is achieved by exploiting the differential thermal expansion coefficients among the layers, where the POE functions as the active layer, while ANFs and CNTs serve as inert reinforcement layers. The bird's-nest-like CNT layer imparts the actuators with repeatable programming capabilities. These intelligent actuators exhibit rapid responses to light, electrical, and thermal stimuli, featuring a low activation energy, high actuation speed, significant deformation, and exceptional fatigue resistance. Inspired by paper cutting and origami techniques, the actuators achieve repeatable morphological programming and complex actuation behaviors. The POE/ANF/CNT composites also demonstrate effective EMI shielding (35.7 dB at 40 wt % CNTs), high tensile strength (39.1 MPa), superior Joule heating performance (301 °C at 20 V voltage), and excellent thermal stabilities (with a maximum decomposition temperature reaching 473 °C). These multifunctional intelligent materials hold significant potential for applications in flexible wearable electronic devices, EMI shielding, and soft robotics.

Moisture-driven actuators are the smart materials or devices that respond to changes in humidity by converting moisture fluctuations into mechanical motion, with versatile applications in soft robotics, bio-inspired systems, smart textiles, and intelligent devices. In this work, a fast and large-deformation moisture-driven Janus fiber actuator with multiscale hydrophilic-hydrophobic interfaces inside is fabricated. Specifically, the Janus fiber actuators are composed of hydrophilic side and hydrophobic side to create a macroscopic hydrophilic-hydrophobic interface, while the highly hydrophilic polyethylene glycol segments in the hydrophilic side form island-like phase domains inside polyurethane matrix and create the microscopic hydrophilic-hydrophobic interfaces. These unique multiscale hydrophilic-hydrophobic interfaces endow the Janus fiber with fast and reversible swelling-deswelling feature, thus enabling it with superior moisture-driven performance. The actuator exhibits a maximum contraction rate of 96.68%, a contraction speed of 58.23% s-1, a curvature of 33.33 cm-1, and a maximum response speed of 14.2 cm-1 s-1, outperforming all reported moisture-driven fiber-based actuators. Due to these exceptional properties, the actuators can be applied in bio-inspired systems to mimic the contraction behaviors of natural actuators. More interestingly, the actuators can also be used as moisture-sensitive smart wigs, which can reversibly switch between short curls and straight hair in response to changes in humidity.

The pursuit of biomimetic fibers with simultaneous high toughness and strength persists, despite their inherent trade-offs.However, for artificial spider silk based on gel fiber, it is still unclear for the molecular chain attributes related to the improvement of the strength and toughness. Here, a hydrogel fiber was prepared by mimicking the molecular structure of natural spider silk, and we delved into the molecular chain structure characteristics related to the strength, toughness and damping capacity of gel fiber, such as crosslinking density, molecular chain orientation and hydrogen bond interaction. The results indicate that a certain increase in crosslinking density and molecular chain orientation contributes to the enhancement of tensile strength, while the toughness and damping remain essentially unaltered. The thermal dissociation of hydrogen bond could enhance the toughness in a specific range, while the humidity destruction of hydrogen bond would reduce the toughness.Through well-regulation control of the weight ratio of polyacrylamide（PAM） to poly（acrylic acid）(PAA), the PAM@PAA gel fibers could reach maximum breaking strength of 1.02 GPa, maximum toughness of 149 MJ m -3 , and damping capacity of 95%.PAM@PAA gel fiber has demonstrated excellent wound healing performance and biocompatibility in vivo evaluation as a surgical suture, which indicates its potential in biomedical applications.

High-performance and moisture-proof piezoelectric nanogenerators (PENGs) are in high demand for pressure sensing and energy harvesting. Inspired by super-hydrophobic structure of lotus leaf, we propose a moisture-proof nanofiber film-based PENG by designing the well-defined microspheres on the surfaces of electrospinning PVDF/ZnO nanorods (NRs) nanofibers, which can long-term serve even under high humidity. More interestingly, attributed to the synergistic effect of the electric field induced crystal transformation of PVDF, stress concentration effect resulting from microspheres and ZnO-NRs, as well as enhanced piezoelectric effect of ZnO-NRs, the as-prepared PENG exhibits various merits, including ultra-high piezoelectric output (103.1 V) with power density of -2.36 mu W cm(-2), fast response time (30.8 ms) and excellent long-term stability over 10,800 cycles. Thanks for the outstanding piezoelectric properties of our PENGs, they can not only be applied in capturing energy such as wind energy, but also employed as flexible wearable sensors to monitor human motions/physiological signals and realize human-machine interaction. This work provides an unexplored strategy for multifunctional flexible piezoelectric sensor with excellent practicability and moisture-proof characteristic.

Liquid crystal elastomers (LCEs) have attracted wide interest due to their characteristic large strain in actuation. However, their high transition temperatures to the isotropic phase (TI) and poor mechanical strength still limit their applications. Here, an LCE network was modified with poly(ethylene glycol) soft segments to lower its TI, which was adjusted close to body temperature. Additionally, a mesh-structured power-net fabric in a prestretched state was introduced into the LCE matrix to reinforce the soft actuators via construction of a double network. The obtained fabric-reinforced LCE exhibited a high Young's modulus (12.3 MPa, 3 times that of LCE) and an actuation strain of 15%. During one actuation cycle, it experienced a retractive stress of 1.21 MPa, accompanied by a maximum work density of 138 kJ/m3. Both the mechanical and actuation properties of the LCE have been improved. This new strategy of combination brought the LCE actuators closer to practical applications.

Self-powered pressure sensors have gained significant attention for their transformative potential in wearable electronics, Internet of Things (IoT) devices, and artificial e-skins. However, attaining high sensitivity while maintaining good breathability has proven to be a formidable challenge. In this study, we design a hierarchically structured all-nanofiber self-powered pressure sensor utilizing the triboelectric and electrostatic induction principles. The sensor is fabricated via an electrospinning process and consists of a multi-layered architecture comprising nanofiber membranes (NMs): a polyvinylidene fluoride/graphene NM as the negative friction layer, an ethyl cellulose/polyvinyl polypyrrolidone NM as the positive friction layer, and silver nanowire-loaded polyurethane NMs as the electrode layers. This innovative all-nanofiber design not only ensures remarkable breathability but also achieves outstanding sensitivity (15.91 V/kPa) and low detection limits (0.0044 N and 1 degrees), attributed to the enhanced surface roughness and amplified surface charge potential of the friction layer. The sensor demonstrates its versatility by accurately monitoring various human motions and performing dual- language character recognition (Chinese and English), highlighting its vast potential for applications in wearable electronics, human-machine interaction, and next-generation e-skins.

ConspectusLiquid crystal elastomer fibers (LCEFs) are reversible artificial muscles capable of stimuli-responsive functions, making them promising competitors for ideal soft actuators. These remarkable actuation properties depend strongly on their mechanical properties, such as elastic modulus and breaking stress. It is necessary to strengthen the LCEF muscles to meet the demands of advanced applications. However, despite the significant progress in LCEFs, there is currently no such Account systematically summarizing and analyzing the strategies adopted for enhancing their mechanical and actuation properties. The intuitive variations among the different enhancement strategies further call for investigations into how to choose the most suitable ones based on specific situations. In this Account, for the first time, we systematically summarize existing approaches to strengthening LCEF-based artificial muscles, contributing to the development of more robust and smarter fibrous artificial muscles.In the first section, we focus on the latest and most valuable progress on strengthening LCEF-based artificial muscles, highlighting the need for a comprehensive summary of the various approaches utilized. The mechanical properties of LCEFs can be tailored through molecular design, physical interactions, and fiber integration. The adjustment of hard/soft segment features, the introduction of additional microstructures, and the fiber integration provide opportunities to strengthen LCEF-based artificial muscles, which are discussed in the second section. Subsequently, we delve into the impact of various preparation methods on the performance of LCEFs, and LCEFs fabricated by different spinning and alignment techniques exhibited rather different mechanical and actuation properties. This has been adopted to engineer novel, stronger, and tailored fibrous artificial muscles, as described in the third section. Moreover, we show that the incorporation of rigid composite materials via coating and doping has emerged as a powerful strategy to strengthen LCEFs, such as core-shell structures. Such enhancements also introduce multifunctionality for LCE-based artificial muscles that can enrich the fiber structure and actuation mechanism, which are elucidated in the fourth section. Finally, we conclude this Account with a critical analysis of the challenges and prospects of LCE-based artificial muscles, hoping to pave the way for the construction of more powerful fibrous artificial muscles.

Integrating electromagnetic wave (EMW) absorption and non-contact sensing into flexible electronic fibers is essential for advancing multifunctional smart textiles. However, their simultaneous realization remains challenging due to the fundamentally opposing requirements for electrical conductivity. Inspired by the regionalized feather structure of an owl's wings, a functionally partitioned core-sheath yarn architecture is designed to spatially decouple EMW absorption and sensing functionalities. The core, composed of polyurethane hybridized with EMW-absorbing nanoparticles that provide magnetic-dielectric synergistic loss mechanisms, exhibits excellent EMW absorption performance (reflection loss peak: -30.1 dB; effective absorption bandwidth: 6.78 GHz). The sheath, made of poly(vinylidene fluoride-trifluoroethylene) nanofibers, offers superior triboelectric charge trapping and hydrophobicity, enabling robust self-powered non-contact sensing. The resulting fabric sensor delivers a high and stable voltage output of 6.3 V at a 2 mm separation, with outstanding durability over 5000 cycles. Integrated with a deep learning model, the sensor array enables real-time gesture recognition, demonstrating its potential in intelligent environmental perception and electromagnetic stealth. This work presents a promising platform for next-generation smart textiles that seamlessly combine energy harvesting, environmental adaptability, and human-machine interaction.

In nature, many animals protect themselves through deformation, discoloration, and infrared concealment to achieve multiple forms of camouflage. Camouflage fabrics designed for deserts and rainforests have vastly different requirements for color, breathability, and infrared emission. However, the development of corresponding smart fabrics remains a significant challenge. In this work, a novel dynamic-bond-controlled hygro-responsive hydrochromic wool fiber artificial muscle, inspired by the octopuses, has been developed, which used to construct a smart multi-camouflage fabric that integrates deformation, allochromasia, and infrared concealment. The obtained fabric exhibits a green color in humid environments (rainforests), with large pores for perspiration. In a dry environment (deserts), the fabric contracts and changes to akhaki color, while the reduction in pore size enhances infrared shielding. As the relative humidity increases from 20% to 100%, the average pore size of multi-camouflage smart fabric decreases by similar to 84%, minimizing the radiative temperature difference between 36 degrees C target. This results in a temperature reduction of 5.2 degrees C for the target. The hydrochromic artificial muscle is also employed to develop smart fabrics that mimic octopus-like behaviors. With its biocompatible, biodegradable, high thermal insulation, and comfortable wearability, the dynamic-bond-based multifunctional muscle fabric opens up additional possibilities for smart textiles, information technology, and artificial intelligence.