Handedness-controlled and solvent-driven actuators with twisted fibers

Bo Fanga, Youhua Xiaoa, Zhen Xu*ab, Dan Changa, Bo Wanga, Weiwei Gaoa and Chao Gao*a
aMOE Key Laboratory of Macromolecular Synthesis and Functionalization, Department of Polymer Science and Engineering, Zhejiang University, No. 38 Zheda Road, Hangzhou 310027, China. E-mail: chaogao@zju.edu.cn
bNational Key Laboratory of Science and Technology on Advanced Composites in Special Environments, Harbin Institute of Technology, Harbin 150080, China. E-mail: zhenxu@zju.edu.cn

Received 24th December 2018 , Accepted 5th March 2019

First published on 6th March 2019

Plenty of biological materials are constructed from repeated unit cells with handed configurations, wherein the hierarchical self-assembly of handed units confers optimized mechanical properties and environmental adaptability to bulk biological materials. Inspired by biological handed architectures, we propose handedness-controlled and solvent-driven actuators by programming twisted fibers, such as twisted graphene oxide fibers (TGFs), with mirrored handedness, mechanical robustness and superb flexibility. The large twists (beyond 4800 turns per meter), hair-like diameter (down to 63 μm), large tensile strain (29%) and light weight (1.49 g cm−3) of TGFs enable them to provide a large start-up torque of 2.7 × 10−7 N m, and to deliver a record rotor kinetic power of 89.3 W kg−1 when stimulated by polar solvents such as acetone and water. By assembling handed TGF units, we achieve precise outputting of rotor kinetic energy (from 0.78 W kg−1 to 12.5 W kg−1), controllable harvesting of electrical energy (from 2.37 W kg−1 to 11.5 W kg−1), and free handling of a heavy object. The activeness, inertness and operation of all the actuating systems are well controlled by the handedness of TGF units. They are highly stable and reversible, and maintain a high energy output efficiency over multiple operation cycles. These handedness-controlled systems are also extended to hybrid twisted fibers containing nanocomposites and polymers, indicating their general practicability. Handedness-controlled actuators open an alternative avenue for fabricating energy harvesters, responsive textiles, electronic skins and soft robots.

Conceptual insights

Handedness-controlled conception has been widely found in nature. Being constructed with repeated unit cells with handed configurations, many biological materials such as actin, collagen, and keratin exhibit optimized mechanical properties and environmental adaptability. The prospects of artificially evolving biological handed architectures for application are promising. But creating this ability in practical engineered systems poses challenges in the design of both reliable unit cells and proper construction. We propose a handedness-controlled rule to construct actuating systems with continuous twisted fibers as the unit cells, with handed helix configurations, mechanical robustness and hair-like diameter. We demonstrate that twisted fibers afford impressive start-up torques driven by polar solvating species, exceeding those of previously reported artificial muscles. The handedness-controlled actuating system controllably outputs rotor kinetic energy in the two-unit system, harvests electrical energy in the three-unit system and handles heavy objects in the four-unit system. The handedness-controlled actuating system delivers power to the load with high precision and efficiency. This handedness-controlled conception is extensively observed for twisted fibers containing graphene oxide, nanocomposites and polymers. This approach is conducive to the design of energy harvesters, responsive textiles, electronic skins and agile soft robots.


A variety of biological materials, such as DNA, actin, collagen, and keratin, are composed of repeated unit cells with handedness. The hierarchical self-assembly of handed units enables bulk materials to exhibit optimized mechanical properties and environmental adaptability.1 For example, fibrous actin (F-actin) composed of right-handed and double-helical globular actin (G-actin) makes muscles to maintain rigidity under tension and flexible under torsion.2 Right-handed triple-helical supercoils assembled from left-handed procollagen helices confer reliable damage tolerance and toughness of tropocollagen, due to the opposite twisting directions of supercoils and procollagen under tension.3 The prospects of artificially evolving biological handed architectures for application are encouraging.4 However, creating this ability in practical engineered systems poses challenges in the design of both reliable unit cells and proper construction at the micron scale.

Inspired by biological handed architectures, we proposed a handedness-controlled rule to design solvent-driven actuators by programming handed twisted fibers, with mirrored helix configurations and hair-like diameter. Some previously reported actuating systems also exhibited helix microstructures, being actuated to transform when exposed to a chemical or physical stimulus.5–14 However, they are limited in several ways. Firstly, practical actuators require diverse actuating models to work in complex operating environments and to deliver the precise amount of energy to meet the actual demand of the receiver, which has rarely been addressed.15 Secondly, the processing of previous helix actuators suffers from some experimental challenges such as tight process control, legacy cost of purification and low yields due to low-efficiency manufacturing facilities.16–18 Thirdly, the energy conversation efficiency and energy output of polymer-19–22 and carbon-based actuators23–27 are still quite low. For example, a polymer/CNT hybrid cloth actuator only afforded a start-up torque of 1.26 × 10−7 N m−1 in response to water,21 and twisted carbon nanotube (CNT) fibers with a rotary stroke of 2050 revolutions per meter yielded a low rotor kinetic power output of 29.7 W kg−1 in the presence of ethanol.24

Here, we report handedness-controlled actuating systems responding to polar solvating species. Starting from low-cost graphene oxide (GO) suspensions, continuous handed TGFs with mechanical robustness and superb flexibility were fabricated by a liquid crystalline wet spinning protocol and an industrial twisting–drawing technique. The prepared TGFs are driven to output rotor kinetic energy by polar solvents selectively, and among them, 0.05 mL acetone drives TGFs to reach a high rotary speed of 6050 rpm, a large start-up torque of 2.7 × 10−7 N m, and a high peak power output of 89.3 W kg−1. These three figures-of-merit reach record high in graphene-based actuators. We programmed multiple TGF units to actuate in a controlled manner. By modulating the handedness of individual TGFs, we precisely achieved rotor kinetic energy from 0.78 W kg−1 to 12.5 W kg−1 in the two-unit system, and electrical energy from 2.37 W kg−1 to 11.5 W kg−1 in the three-unit system and handled a heavy object beyond 60 times its own weight in the four-unit system. This handedness-controlled approach can meet the actual demand of practical applications and is also conducive to the design of responsive textiles, electronic skins and agile soft robots. The general practicability of this handedness-controlled conception is demonstrated by our investigation of hybrid twisted fibers containing nanocomposites and polymers, such as carbon nanotubes (CNTs), polyvinyl alcohol (PVA), and titania nanoparticles.

Results and discussion

Derived from traditional wet-spinning protocols,28–32 continuous GO belts were first obtained by a liquid crystalline wet spinning method and the continuous extrusion of dopes was controlled using a micro-fluidic system (Fig. 1a). In particular, GO liquid crystalline dispersions in dimethyl formamide (DMF, 10 mg mL−1) were used as spinning dopes, then extruded into ethyl acetate (EA) through a micro-fluidic channel at a rate of 20 m min−1. The microfluidic system was a hollow gallery with a width of 2 mm and a height of 200 μm. Solvent exchange and hydrogen-bond interactions facilitated the formation of gel belts in EA, and the birefringence shown in Fig. 1b indicated the oriented alignment of GO sheets in the gel belts aided by shear flowing. With the rapid removal of DMF and EA in air, gel belts evolved into neat GO belts with a width of 200 μm and favorable stretchability (Fig. 1c and Fig. S1, ESI).
image file: c8mh01647j-f1.tif
Fig. 1 The fabrication of continuous and robust TGFs with mirrored handedness. (a) Optical image of continuous GO belts passing through a micro-fluidic channel. (b) POM image of the as-prepared GO belts. (c) SEM image of knotted GO belts. The picture (d) and scheme (e) of a drawing–twisting process to obtain continuous TGFs. SEM images of a RTGF (f and g) and a LTGF (h and i), and the cross-sectional observation (j). (k) The relationship between the helical angle, inserted twist and diameters.

Flexible GO belts were processed into continuous TGFs using a twisting–drawing apparatus (Fig. 1d). The out-of-plane revolution of the roller generated a normal torque (Fig. 1e), managing to twist axially oriented GO belts (Fig. S2a and b, ESI) into spirally oriented TGFs (Fig. S2e and f, ESI), and further storing strain energy into TGFs. Simultaneously, the in-plane rotation of the roller functioned to collect TGFs at a rate of 7 cm min−1. A right-handed TGF (RTGF, Fig. 1f and g) and a left-handed TGF (LTGF, Fig. 1h and i) with a smooth surface and uniform circular cross section (Fig. 1j) were acquired by adjusting the rotation direction. Fig. 1k displays the diameter (D) dependence of the helical angle (θ, degrees) with respect to the twisting direction and the twist number in turns per meter (T = tan[thin space (1/6-em)]θ/(πD), turns per m).33,34 Both θ and T increase with decreasing D, and T reaches 4880 turns per m when D decreases to 63 μm (Fig. S3, ESI). It is noteworthy that our method provides a new type of protocol to achieve continuous TGFs,9–11 saving the trouble of batch-to-batch disparity and low productivity. Continuous TGFs exhibit favorable flexibility and mechanical robustness, as the large inserted twist evokes a great capability to deform without tearing. A typical tensile stress–strain curve for a 63 μm-thick TGF shows a large tensile strength of 130 MPa. The extreme elongation prior to failure is 29% (Fig. S4, ESI), which is around 10 fold that (<3%) of highly stretched GO fibers.35–39

Benefitting from their superb mechanical durability (see the discussion in Fig. S4, ESI), TGFs output rich rotor kinetic energy from solvating stimulators. We investigated the actuating behavior of a suspended LTGF (∼40 μg) with a length of 20 cm, which was loaded with a copper paddle weighing 50 mg (Fig. 2a and Movie S1, ESI). After feeding 0.05 mL acetone, the copper paddle was accelerated to an angular speed (ω) of 633 rad s−1 in a short period of 0.7 s, with an average acceleration of 904 rad s−2. Since the moment of inertia (J) is measured as 3 × 10−10 kg m−2, the maximum start-up torque (τ) is calculated as 2.7 × 10−7 N m, which is 18 times that of a moisture-driven graphene fiber motor,25 3.6 times that of a CNT fiber actuator,24 and 2.1 times that of a polymer-based fiber actuator.21 Such a large start-up torque provides a high peak power output (p = 2/2t) of 89.3 W kg−1 to the copper paddle, superior to CNT artificial muscles40 (61 W kg−1) and a graphene oxide fiber motor (71.9 W kg−1). The revolutions per minute (rpm) are 6050, which is the highest value ever achieved by graphene-based kinetic energy actuators,25,41 and much higher than those of polymer-based actuating systems.19–21

image file: c8mh01647j-f2.tif
Fig. 2 Handedness-controlled rotor kinetic energy harvesters constructed with solvent-driven TGFs. (a) The optical images of the reversible rotation of a TGF in the presence of 0.05 mL acetone. The relationship of rotary speed (b), start-up torque and peak power output (c) with the polarity of exposed solvents. (d) The measured speeds of the forward rotation and reverse rotation of TGFs driven by polar solvating species for 100-cycle tests. The scheme recording the responses of two united TGFs suffering the successive wetting of acetone, with the alternating handedness of RR (e) and RL (f). (g) The scheme showing that the actuation of a two-TGF system is realized by their resultant torque. (h) The relationship between the peak power output, rotary speed and spacing width between two RR or LL units. (i) The monolayer fabric woven with cotton fibers and a LL configuration. The composite fabric formed a helical morphology when wetted in an acetone bath.

In particular, the actuating properties of a TGF rotor are well controlled by adjusting the polarity of solvating species. We tested some solvents with an individual volume of 0.05 mL and increasing polarity (Fig. 2b and c). In general, high peak output powers of 39.7, 33.9, 29.7 and 27 W kg−1 were achieved upon exposing TGFs to highly polar stimulators, e.g., methanol, water, ethanol and DMSO, and the start-up torques exceeded those of previously reported CNT fiber-based artificial muscles (ref. 15, 16, and 35 in Fig. 2c). Inefficient energy output values were recorded for other polar solvents (IPA, EA, TCM, DCM and DMF), which yielded peak powers between 0.1 and 20 W kg−1. By contrast, stimulation by non-polar solvents was fairly weak. The rotary speeds for hexane, cyclohexane and methylbenzene are 88, 79 and 49 rpm, respectively, and the system reached inertness in the case of 1-octanol.

The strong actuating ability of TGFs driven by polar solvents can be explained in two aspects. In one case, the oxygen-rich functional groups on the surfaces of TGFs induce good affinity and rapid infiltration of polar molecules,42–44 thus resulting in fast radial expansion and untwisting of TGFs. This explanation is verified by the diffusion speed difference of solvents on TGFs observed by confocal laser scanning microscopy (CLSM, Fig. S6, ESI). Uniform diffusion of acetone along the axial direction was clearly observed for 4 seconds, where the diffusion speed of acetone was faster than those of other solvating species. In the other case, during the infiltration of polar solvents, a strong surface tension (γ, the inset in Fig. 2b) and a contact angle (α) would increase along the liquid–solid interface.45 Therefore, an elastocapillary force (Fe = γ[thin space (1/6-em)]sin[thin space (1/6-em)]α) vertical to the interface was triggered, initiating the rapid untwisting. While pure nonpolar solvents were inert, they can be applied to adjust their apparent actuating properties. In a case study of mixing 1-octanol (inert) and acetone (active, Fig. S7a, ESI), the measured rotary speed increased gradually upon addition of acetone, leading to a peak output power beyond 70 W kg−1 as the volume fraction of acetone was beyond 70%.

The torsional rotation of TGFs was demonstrated to be repeatable at least 100 times when triggered by some polar solvents, such as acetone, methanol and ethanol (Fig. 2d). As discussed above, the infiltration of polar solvents induced the untwisting of TGFs. To guarantee the reversibility of torsional rotation, we tethered the TGF at both ends to prohibit the excessive release of strain energy. Highly volatile fluids, i.e. acetone, methanol and ethanol, departed from TGFs soon after the untwisting finished, causing negligible harm to helical microstructures. The rapid release of strain energy in the untwisting process produced a reverse internal stress, which drove untwisted TGFs to rotate reversely to rebuild the initial architectures. The reverse rotation was slower than the forward rotation. For example, the recovery speed of TGFs decreased to 4830 rpm from a forward speed of 6050 rpm when actuated by acetone, to 3510 rpm from 4200 rpm when driven by methanol and to 1800 rpm from 3300 rpm when stimulated by methanol and ethanol. The reversibility and repeatability of liquid water were poor, since liquid water swelled GO laminates, destroying the helical structures of TGFs in the first cycle. Due to their non-volatile properties, high viscosity fluids, such as DMSO, also suppressed the reversibility and repeatability of actuating systems.

Uniting a couple of units into a handedness-controlled actuating system can precisely control the rotary speed and kinetic energy output. The configuration of the two TGF system is either homochiral (RR or LL) or heterochiral (RL or the equivalent LR) depending on the relative handedness. For the RR configuration, the system remained silent when one unit was wetted with acetone, but the RR system coiled rapidly (see Movie S2, ESI) after the other unit was wetted (Fig. 2e). This coiled configuration is highly similar to the architecture of F-actin.2 The LL system showed similar behavior to the RR system but rotated toward the reverse direction. The SEM images in Fig. S7b (ESI) indicated that the screw pitch of a coiled helix switched between the millimeter level (1.7 mm) and the micron level (320 μm). When we replaced a L unit with a R unit, the built RL/LR configuration was immune to the consecutive doping with stimulating solvents (Fig. 2f). The two torques generated in RR/LL configurations (τ1 and τ2, respectively, Fig. 2g) had the same magnitude and direction, contributing equally to the torsional rotation. While τ1 and τ2 completely cancelled each other in the RL/LR configurations due to the opposite direction, the rotary speed of the RR/LL configurations was modulated by tailoring the spacing width (W) between the two fibers (Fig. 2h). By increasing the value of W from 2 mm to 10 mm, we demonstrated that the rotary speed decreased from 3190 rpm to 800 rpm in the RR system, accompanied by a decrease of power output from 12.5 W kg−1 to 0.78 W kg−1, while the corresponding power output in the LL system decreased from 12.3 W kg−1 to 0.77 W kg−1. The above tests further inspire us to create a ring-like actuator (Fig. S8a, ESI) and a smart textile (Fig. 2i and Fig. S8b–e, ESI), which exhibited repeatable actuating motion upon repeated exposures to acetone.

Based on the optimization studies, we adapted handedness-controlled actuating systems to collect electrical energy from the rotor kinetic energy by modulating the handedness of three units. Uniting three units can activate a high moment-of-inertia magnet rotor, thus improving the energy harvesting efficiency. More units are not suitable due to the lower rotary speed (<5 rpm, as discussed later). The electrical power output frequency changed from 10 Hz to 5 Hz in three-TGF-based electromagnetic induction systems, i.e., RRR, LLL, RRL, RLL and RLR (Fig. 3a). For the RRR configuration, the rotation generated a peak gravimetric electrical power of 9.8 W kg−1, corresponding to an output electrical energy of 0.86 J kg−1 per cycle (Fig. 3b). Because the rotary speed was 600 rpm, the peak power output was calculated to be 18.15 W kg−1. Thus, the kinetic-to-electrical energy conversion coefficient was 54%, superior to those of artificial polymer muscles.46 Similarly, the LLL configuration achieved a gravimetric electrical power of 11.5 W kg−1, an output electrical energy of 0.88 J kg−1 per cycle and an energy conversion coefficient of 57%. For the RRL configuration, the reverse torque produced by the L unit in the end partially counteracted the middle R unit, providing a gravimetric output electrical power of 5.9 W kg−1 and an electrical energy per cycle of 0.51 J kg−1, around 60% and 58% of those of the RRR configuration, respectively. The gravimetric output electrical power and energy per cycle by the RLR configuration decreased to 2.37 W kg−1 (23% of the RRR configuration) and 0.21 J kg−1 (24% of the RRR configuration), since the torque of the middle LTGF cancelled those of the remaining RTGFs. All the configurations maintained high and stable energy conversion efficiencies from 50% to 60% during our cycle tests (Fig. 3c). The above tests demonstrate the feasibility of three-TGF systems to precisely control the electrical energy harvesting by programming the handedness of unit cells, while maintaining high efficiency and stability.

image file: c8mh01647j-f3.tif
Fig. 3 Handedness-controlled electrical energy harvesters constructed with RRR, LLL, RRL, RLL and RLR configurations, respectively. The dependences of gravimetric output electrical power (a) and electrical energy (b) on the handedness and arrangement of TGF units. (c) The kinetic-to-electrical energy conversion coefficients of all the configuration during 10 cycle tests.

We modified a handedness-controlled actuating system to operate as a soft actuator, which handled a heavy object in a controlled mode. This is an ability indispensable for soft robotics. The soft actuator was constructed by tethering four TGF units with a pendant weighing 2.5 mg (m′), around 61 fold that of a single unit (41.7 μg). For the RRRR system, the united axial extension of the four TGFs caused the pendant to experience a helical swing when actuated by in-plane torques. The pendant rose from the horizontal position to a nearly vertical position for 15 seconds (t′) (Fig. 4a) with the increase of the tilting angle from 0° to 88° and a constant swing speed of 5.87° s−1 (Fig. S9, ESI). The mass center was raised by 1.2 cm (h). Thus, the output gravitational potential energy (G = mgh) and gravitational potential power (Pg = G/t′) were calculated to be 17.66 J kg−1 and 1.17 W kg−1, respectively. Due to the clockwise synergistic torque produced by the four RTGFs, the RRRR-TGF system swung in a clockwise direction. As a comparison, the pendant in the LLLL system swung to the anticlockwise direction and a height of 1.2 cm at a speed of 5.90° s−1 (Fig. S10, ESI). The nearly same swing speed and rising height of the RRRR configuration as those of the LLLL configuration suggests the reliable symmetry of the handedness-controlled actuating system (Fig. S9, ESI). For LRRR, the system swung in the unaltered direction, but the rising height and G values decreased to 0.76 cm and 11.2 J kg−1 (Fig. 4b and Movie S3, ESI). This is because the new added LTGF produced an anticlockwise torque counteractive to those of the remaining three RTGFs, weakening the synergistic torsion action. The rising height and G were decreased to 0.48 cm and 7.04 J kg−1 in the RLRR configuration (Fig. 4c), due to the bilateral influence of the middle torque of the LTGF, as discussed in the RLR configuration (Fig. 3a).

image file: c8mh01647j-f4.tif
Fig. 4 Handling a heavy object using a four-TGF actuating system. The pictures in (a), (b) and (c) recording the movement of the heavy object in RRRR, LRRR, and RLRR configurations driven by acetone, which exhibits controllable swing and lifting. The relationship of the swing direction (d), lifting heights and output gravitational energy (e) of the actuating system with the helical chirality of the components.

By adjusting the arrangement of units in four-TGF configurations, we conclude that the working direction is controlled by the quantitative proportion of two kinds of units, which is nothing to do with the spatial arrangement. As displayed in the red zone of Fig. 4d, RRRR, RRRL, RRLR, RLRR, and LRRR configurations occurred in the clockwise swing since the produced torques of the RTGF overmatched those of the LTGF. And the blue zone represented the anticlockwise swing where LTGF units predominated. The system remained silent when RTGF units and LTGF units were numerically equal (e.g. RLLR, RLRL and RRLL configurations), because they counteracted each other. Different from the working directions, the gravitational energies and power outputs are closely related to the spatial arrangement of units. We summarized the time dependences of the rising height and average gravitational energy output of the RTGF-dominated configurations (Fig. 4e), finding that the movement of the suspended object maintained a constant speed in all the systems. In the whole working process, the G remained at 17.66 J kg−1 in the RRRR configuration, and decreased by 40% in the LRRR configuration, and by 35% in the RLRR configuration.

All the handedness-controlled systems discussed above exhibit superb controllability and high efficiency. More importantly, this phenomenon can also be extended to other unit cells by hybridizing different materials with TGFs. We have shown different examples of hybrid TGFs with CNTs, PVA, and titania nanoparticles, individually. Nanomaterials with weight fractions beyond 50% dispersed uniformly in the hybrid fibers (Fig. 5a and Fig. S11, ESI). The operation capacities of twisted fibers fabricated using different materials (i.e. graphene oxide, CNTs, PVA and titania) were mainly decided by four factors. They are the hydrophily of materials, the affinity between materials and fluid stimuli, the torsion moduli and the twist numbers in turns per meter (T). Carboxylated CNTs and PVA are more hydrophilic than titania nanoparticles, so twisted fibers containing CNTs and PVA show better affinity with polar solvents than hybrid fibers containing titania particles. Driven by 0.05 mL acetone, R-type titania fibers outputted a lower rotor kinetic power (24.3 W kg−1) than those of R-type CNT fibers (45.1 W kg−1) and R-type PVA fibers (26.8 W kg−1). Experimentally, CNTs are easily liquid-phase processed, while PVA is difficult to twist due to its low torsion modulus. Thus, twisted fibers containing CNTs own larger T than PVA fibers, which can be directly determined from Fig. 5a and Fig. S11 (ESI). The larger T contributes to the higher rotor kinetic power output of CNT fibers than that of PVA fibers. The rotor kinetic power outputs of the hybrid fibers are superior to those of previously reported artificial muscles13,15 (Fig. 5b). The nearly equal energy outputs of R-type fibers and L-type fibers reflected a favorable operating symmetry. In two-unit systems, RR- or LL-type actuators delivered rotor kinetic power varying from 9.8 W kg−1 to 5.3 W kg−1. RL- or LR-type systems remained inert, since the two opposite torques used in the rotor completely cancelled each other. Similar to the three-unit systems of neat TGFs, nanocomposite-based systems outputted highly controllable power by adjusting the arrangement of cell units with different handednesses. In a case study of CNTs, the power outputs of RRR-, RRL- and RLR-systems showed a gradual decrease from 7.9 W kg−1 to 4.1 W kg−1. Through the study on handed nanocomposite-based hybrid fibers, we demonstrated the general applicability of handedness-controlled conception in practical actuating systems.

image file: c8mh01647j-f5.tif
Fig. 5 (a) Hybrid twisted fibers made with graphene oxide and carbon nanotubes, polyvinyl alcohol, and titania nanoparticles. (b) The power outputs of the hybrid twisted fibers in handedness-controlled actuating systems in the presence of 0.05 mL acetone.


In summary, we designed handedness-controlled actuating systems by programming flexible handed TGFs. Individual TGFs exhibit high stretchability and mechanical robustness, giving rise to an efficient kinetic energy output driven by various polar solvents. As unit cells, TGFs are logically programmed into intricate actuating systems, which operate as highly controllable soft actuators to handle a heavy object and harvest energy with a high energy conversion efficiency. The high controllability, easy operation, and flexibility of TGFs allow them to be used as high-precision energy harvesters and soft robots. We widely apply handedness-controlled systems to other unit cells, such as hybrid twisted fibers containing nanocomposites and polymers.

Conflicts of interest

The authors declare no conflicts of interest.


The authors acknowledge the kind supply of massive giant GO spinning dopes from the company Gaoxi Tech (http://www.gaoxitech.com). This work was supported by the National Natural Science Foundation of China (Grant No. 51533008, 51703194, 51603183, and 21325417), the National Key R&D Program of China (Grant No. 2016YFA0200200), the Fundamental Research Funds for the Central Universities (Grant No. 2017XZZX008-06), and the Hundred Talents Program of Zhejiang University (188020*194231701/113).

Notes and references

  1. E. Renuart and C. Viney, Pergamon Mater. Ser., 2000, 4, 223–225 CrossRef.
  2. H. Lodish, D. Baltimore, A. Berk, S. L. Zipursky, P. Matsudaira and J. Darnell, Mol. Cell Biol., W. H. Freeman and Company, New York, 1995 Search PubMed.
  3. C. A. Smith and E. J. Wood, Biol. Mol., Chapman and Hall, London, 1991 Search PubMed.
  4. J. I. Lipton, R. MacCurdy, Z. Manchester, L. Chin, D. Cellucci and D. Rus, Science, 2018, 360, 632–635 CrossRef CAS PubMed.
  5. D. Li, W. F. Paxton, R. H. Baughman, T. J. Huang, J. F. Stoddart and P. S. Weiss, MRS Bull., 2009, 34, 671–681 CrossRef CAS.
  6. J. E. Huber, N. A. Fleck and M. F. Ashby, Proc. R. Soc. London, Ser. A, 1997, 453, 2185–2205 CrossRef.
  7. J. Foroughi, G. M. Spinks, G. G. Wallace, J. Oh, M. E. Kozlov, S. Fang, T. Mirfakhrai, J. D. W. Madden, M. K. Shin, S. J. Kim and R. H. Baughman, Science, 2011, 334, 494–497 CrossRef CAS PubMed.
  8. H. T. Guo and W. H. Liao, Smart Mater. Struct., 2012, 21, 065012 CrossRef.
  9. D. Hwang and T. Higuchi, IEEE-ASME Trans. Mech., 2014, 19, 1625–1635 Search PubMed.
  10. J. Yao, Z. Jiao, D. Ma and L. Yan, IEEE-ASME Trans. Mech., 2014, 19, 633–641 Search PubMed.
  11. J. Li, H. Zhao, H. Qu, T. Cui, L. Fu, H. Huang, L. Ren and Z. Fan, Sens. Actuators, A, 2013, 194, 269–279 CrossRef CAS.
  12. M. D. Lima, N. Li, M. J. Andrade, S. Fang, J. Oh, G. M. Spinks, M. E. Kozlov, C. S. Haines, D. Suh, J. Foroughi, S. J. Kim, Y. S. Chen, T. Ware, M. K. Shin, L. D. Machado, A. F. Fonseca, J. D. W. Madden, W. E. Voit, D. S. Galvao and R. H. Baughman, Science, 2012, 338, 928–932 CrossRef PubMed.
  13. P. Chen, S. He, Y. Xu, X. Sun and H. Peng, Adv. Mater., 2015, 27, 4982–4988 CrossRef CAS PubMed.
  14. S. Kim, C. S. Haines, N. Li, K. Kim, T. Mun, C. Choi, J. Di, Y. Oh, J. Oviedo, J. Bykova, S. Fang, N. Jiang, Z. Liu, R. Wang, P. Kumar, R. Qiao, S. Priya, K. Cho, M. Kim, M. Lucas, L. F. Drummy, B. Maruyama, D. Lee, X. Lepro, E. Gao, D. Albarq, R. O. Ronles, S. Kim and R. H. Baughman, Science, 2017, 357, 773–778 CrossRef CAS PubMed.
  15. S. Li, D. M. Vogt, D. Rus and R. J. Wood, Proc. Natl. Acad. Sci. U. S. A., 2017, 201713450 Search PubMed.
  16. M. Zhang, K. R. Atkinson and R. H. Baughman, Science, 2004, 306, 1358–1361 CrossRef CAS PubMed.
  17. R. C-Silva, A. M-Gomez, H. Kim, H. Jang, F. Tristan, S. V-Diaz, L. P. Rajukumar, A. Elias, N. P-Lopez, J. Suhr, M. Endo and M. Terrones, ACS Nano, 2014, 8, 5959–5967 CrossRef PubMed.
  18. X. Xiang, Z. Yang, J. Di, W. Zhang, R. Li, L. Kang, Y. Zhang, H. Zhang and Q. Li, Nanoscale, 2017, 9, 11523–11529 RSC.
  19. S. Jiang, F. Liu, A. Lerch, L. Ionov and S. Agarwal, Adv. Mater., 2015, 27, 4865–4870 CrossRef CAS PubMed.
  20. L. Zhang, H. Liang, J. Jacob and P. Naumov, Nat. Commun., 2015, 6, 7429 CrossRef CAS PubMed.
  21. J. Gong, H. Lin, J. W. C. Dunlop and J. Yuan, Adv. Mater., 2017, 29, 1605103 CrossRef PubMed.
  22. C. S. Haines, M. D. Lima, N. Li, G. M. Spinks, J. Foroughi, J. D. W. Madden, S. H. Kim, S. Fang, M. J. d. Andrade, F. Goktepe, O. Goktepe, S. M. Mirvakili, S. Naficy, X. Lepro, J. Oh, M. E. Kozlov, S. J. Kim, X. Xu, B. J. Swedlove, G. G. Wallace and R. H. Baughman, Science, 2014, 343, 868–872 CrossRef CAS PubMed.
  23. M. F. L. De Volder, S. H. Tawfick, R. H. Baughman and A. J. Hart, Science, 2013, 339, 535–539 CrossRef CAS PubMed.
  24. P. Chen, Y. Xu, S. He, X. Sun, S. Pan, J. Deng, D. Chen and H. Peng, Nat. Nanotechnol., 2015, 10, 1077–1083 CrossRef CAS PubMed.
  25. H. Cheng, Y. Hu, F. Zhao, Z. Dong, Y. Wang, N. Chen, Z. Zhang and L. Qu, Adv. Mater., 2014, 26, 2909–2913 CrossRef CAS PubMed.
  26. Q. Shi, J. Li, C. Hou, Y. Shao, Q. Zhang, Y. Li and H. Wang, Chem. Commun., 2017, 53, 11118–11121 RSC.
  27. C. Hua, Y. Shang, X. Li, X. Hu, Y. Wang, X. Wang, Y. Zhang, X. Li, H. Duan and A. Cao, Nanoscale, 2016, 8, 10659–10668 RSC.
  28. Z. Xu and C. Gao, Nat. Commun., 2011, 2, 571 CrossRef PubMed.
  29. B. Fang, L. Peng, Z. Xu and C. Gao, ACS Nano, 2015, 9, 5214–5222 CrossRef CAS PubMed.
  30. Y. Liu, Z. Xu, W. Gao, Z. Cheng and C. Gao, Adv. Mater., 2017, 29, 1606794 CrossRef PubMed.
  31. L. Kou, Y. Liu, C. Zhang, L. Shao, Z. Tian, Z. Deng and C. Gao, Nano-Micro Lett., 2017, 9, 51 CrossRef PubMed.
  32. B. Fang, J. Xi, Y. Liu, F. Guo, Z. Xu, W. Gao, D. Guo, P. Li and C. Gao, Nanoscale, 2017, 140, 12178–12188 RSC.
  33. K. Chun, S. Kim, M. Shin, C. Kwon, J. Park, Y. Kim, G. M. Spinks, M. D. Lima, C. S. Haines, R. H. Baughman and S. Kim, Nat. Commun., 2014, 5 Search PubMed.
  34. C. S. Haines, N. Li, G. M. Spinks, A. E. Aliev, J. Di and R. H. Baughman, Proc. Natl. Acad. Sci. U. S. A., 2016, 113, 11709–11716 CrossRef CAS PubMed.
  35. Z. L. Dong, C. C. Jiang, H. H. Cheng, Y. Zhao, G. Q. Shi, L. Jiang and L. T. Qu, Adv. Mater., 2012, 24, 1856–1861 CrossRef CAS PubMed.
  36. Z. Xu and C. Gao, Acc. Chem. Res., 2014, 47, 1267–1276 CrossRef CAS PubMed.
  37. H. Cheng, C. Hu, Y. Zhao and L. Qu, NPG Asia Mater., 2014, 6, e113 CrossRef CAS.
  38. Z. Li, Z. Liu, H. Sun and C. Gao, Chem. Rev., 2015, 115, 7046–7117 CrossRef CAS PubMed.
  39. Z. Xu and C. Gao, Mater. Today, 2015, 18, 480–492 CrossRef CAS.
  40. J. Foroughi, G. M. Spinks, G. G. Wallace, J. Oh, M. E. Kozlov, S. Fang, T. Mirfakhrai, J. D. W. Madden, M. K. Shin, S. J. Kim and R. H. Baughman, Science, 2011, 334, 494–497 CrossRef CAS PubMed.
  41. S. Park, J. An, J. Suk and R. S. Rouff, Small, 2010, 6, 210–212 CrossRef CAS PubMed.
  42. S. Park, J. An, I. Jung, R. D. Piner, S. An, X. Li, A. Velamakanni and R. S. Rouff, Nano Lett., 2009, 9, 1593–1597 CrossRef CAS PubMed.
  43. D. Kim, Y. Yun and H. Jin, Curr. Appl. Phys., 2012, 12, 637–642 CrossRef.
  44. D. Konios, M. M. Stylianakis, E. Stratakis and E. Kymakis, J. Colloid Interface Sci., 2014, 430, 108–116 CrossRef CAS PubMed.
  45. P. Wang, R. Bian, Q. Meng, H. Liu and L. Jiang, Adv. Mater., 2017, 29, 1703042 CrossRef PubMed.
  46. S. H. Kim, M. D. Lima, M. E. Kozlov, C. S. Haines, G. M. Spinks., S. Aziz, C. Choi, H. J. Sim, X. Wang, H. Lu, D. Qian, J. D. Madden, R. H. Baughman and S. J. Kim, Energy Environ. Sci., 2015, 8, 3336–3344 RSC.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c8mh01647j

This journal is © The Royal Society of Chemistry 2019