Recent advances in integration of 2D materials with soft matter for multifunctional robotic materials

Abstract

Emerging soft robots with infinite degrees of freedom are one step closer to providing better human–machine interactions than conventional hard and stiff robots, attributing to their outstanding compliance/adaptability, evenly distributed stress, and programmable actuating behaviors. Soft matter (e.g., hydrogels and elastomers) with high mechanical stability has been usually adopted for the fabrication of soft robots. However, soft matter exhibits limited optical, electrical, thermal, and chemical properties that restrict the development of functional soft robots. An emerging approach is to develop multifunctional robotic materials that are reconfigurable and can provide diverse built-in functions, such as wide-spectrum protection, tactile sensing, remote control, and wireless communication. To realize this approach, two-dimensional (2D) materials with diverse yet unique physicochemical properties have been recently integrated with soft matter to bring in add-on functionalities for fabricated soft robots. In this Minireview, we highlight three integration approaches for the fabrication of 2D material–soft matter robotic materials: (i) heterogenous blending of 2D materials within soft matter precursors followed by in situ crosslinking/curing; (ii) bilayer integration of 2D materials with soft matter substrates; and (iii) post-stabilization of 2D material (or 2D material-templated) architectures with elastomers. The advantages and drawbacks of each approach regarding the fabrication process and resulting characteristics are discussed in detail. The reversible actuating behaviors and built-in capabilities of the as-fabricated 2D material–soft matter composites, as well as their further applications as multifunctional robotic materials are summarized. Finally, current research gaps and future directions regarding the development of multifunctional robotic materials are addressed from our perspective by considering the design principles for future untethered soft robots.

---

Lin Jing

Lin Jing received his PhD in Materials Science and Engineering from Nanyang Technological University, Singapore, in 2019. He is currently a research fellow under Prof. Po-Yen Chen in the Department of Chemical and Biomolecular Engineering at the National University of Singapore. His research interests lie in the topographical engineering of 2D materials and their applications in soft electromechanical and energy harvesting devices.

Kerui Li

Kerui Li is a postdoctoral research fellow at the National University of Singapore under the supervision of Prof. Po-Yen Chen. He received his PhD in Material Science and Engineering from Donghua University, Shanghai, in 2017. Currently, he is focusing on light modulation/conversion from solar light to infrared radiation for smart clothing and soft robotics.

Haitao Yang

Haitao Yang received his bachelor's and master's degrees in Materials Science and Engineering from Zhengzhou University. Now he is a PhD candidate under the supervision of Prof. Po-Yen Chen in the National University of Singapore. His research interest focuses on developing multifunctional backbone materials for soft robotics.

Po-Yen Chen

Po-Yen Chen is currently an assistant professor in the Department of Chemical and Biomolecular Engineering at the National University of Singapore (NUS). He completed his PhD in Chemical Engineering from the Massachusetts Institute of Technology (MIT) and was awarded a Hibbitt Independent Research Fellowship at Brown University. He also received an AME Young Investigator Award at Singapore. His research team focuses on strain engineering in 2D materials for the creation of mechanically-stable structures that can sustain large deformations and still preserve the intrinsic functionalities of 2D materials. Based on these insights, his team aims to develop stretchable 2D-material electronics for wearable technologies and multifunctional robotic materials for smart soft robots.

---

Introduction

Emerging soft robots with infinite degrees of freedom are one step closer to providing better human–machine interactions than conventional hard and stiff robots, attributed to their outstanding compliance, excellent adaptability, complex motions, evenly distributed force, and programmable actuating behaviors.^1–5 These merits promise necessary safety for both humans and robots due to highly compliant contacts and less concentrated stress achieved through low-stiffness robotic bodies. Moreover, soft robots generally feature light weight, low fabrication costs, efficient power consumption, and easy accessibility.^6–8 To date, soft robots have been applied to different fields ranging from pick-and-place operations,^9,10 to artificial muscles,^5,11,12 prostheses,^13,14 drug-delivery devices,^15,16 and surgical robotic arms.^15,17,18

Most of the earlier-reported soft robots were fabricated using soft matter with low Young's moduli, large deformability, and high mechanical stability, such as hydrogels and elastomers. These soft robots demonstrate quick changes in material properties (e.g., stiffness) in response to certain stimuli.^1,7,19 These soft matter robotic bodies exhibit sufficient mechanical stability and deformability, but exhibit limited optical, electrical, thermal, and chemical properties, severely restricting the development of robotic functions in sensation, protection, communication, and programmability. As such, to incorporate add-on functionalities into soft robots, conventional strategies for equipping external electronics, such as new modules^20–22 and exoskeletons,^23,24 have been adopted and investigated. Nevertheless, this approach requires a high level of system integration and inevitably increases the overall weight of robotic systems, which contradicts the ultimate goal of creating light, compliant, power-efficient, and eventually untethered soft robots. Therefore, instead of equipping external electronic modules/devices, an alternative strategy is to develop multifunctional robotic materials that can meet two general requirements, reconfigurability and multifunctionality. Reconfigurability denotes the shape or structural recoverability of specific (meta-)materials,^25,26 which are critical for fabricated soft robots to undergo dynamic morphological changes and quick reconfigurations. Multifunctionality indicates the augmented functions of soft robotic materials on top of necessary mechanical stability,⁷ such as threat protection, environmental mediation, mechanical/chemical sensing, energy harvesting, and even wireless communication.²⁷

To achieve the aforementioned candidate robotic materials, one promising route is to integrate functional nanomaterials with soft matter to achieve required multifunctionality and reconfigurability. Recently, two-dimensional (2D) materials, such as graphene and graphene oxide (GO),^27,28 hexagonal boron nitride (h-BN),^29–31 transition metal carbides (MXenes),^32,33 transition metal dichalcogenides (TMDs),³⁴ and black phosphorus (BP),³⁵ have aroused enormous research interest because of their superior electrical, mechanical, optical, and barrier properties, which have been widely applied to sensors,^18,35 actuators,^36,37 and barrier technologies.^38,39 These unique physicochemical properties of 2D materials are highly desired for soft robots, enabling them to work in different environments and respond in real-time for achieving optimal human–robot and robot–robot interactions. Nevertheless, the incompatibility between “hard” 2D materials and “soft” matter has been a huge challenge towards full utilization of their intrinsic physicochemical properties in fabricated soft robots.^40,41

In this Minireview, we focus on three approaches to integrating a wide range of 2D materials with various soft matter (Fig. 1), including (i) heterogeneous blending, (ii) bilayer integration, and (iii) post-stabilization. Meanwhile, we summarized the augmented functions of the resulting 2D material–soft matter composites, and their potential in serving as new robotic bodies/skin with built-in multifunctionality is further reviewed. First, a simple and straightforward method is introduced by heterogeneously blending 2D materials into soft matter to obtain various 2D material–soft matter composites, which can sustain large and reversible deformations and simultaneously provide tactile sensing capabilities during robotic actuations. The heterogeneously blended 2D material–soft matter composites further lead to the fabrication of remotely controlled soft robots driven by light. Second, to fully utilize the unique functionalities of 2D materials under stretching and large deformations, we summarize the studies reporting bilayer integration of 2D materials with soft matter substrates, the applications of which in stretchable protection, stretchable tactile sensing, bio-encapsulation and sensing, as well as wireless communication, are discussed. Third, the post-stabilization approach is reviewed, where 2D material assemblies or 2D material-templated architectures are infiltrated with thin soft matter, enabling their potential applications in physical protection, durable tactile sensation, and multifunctional robotic backbones for tactile sensing and wireless communications. Finally, current research gaps and future directions regarding the development of multifunctional robotic materials at multiple length scales are addressed from our perspectives.

Fig. 1 Multiple approaches to achieving 2D material–soft matter robotic materials with multifunctionality. Schematic illustration of integrating various 2D materials (e.g., graphene, MXenes, BP, h-BN, TMDs, adapted from ref. 38) with soft matter (e.g., elastomers, hydrogels). Diverse approaches are developed, including heterogeneous blending, bilayer integration, and post-stabilization. The as-fabricated 2D material–soft matter composites can serve as multifunctional robotic materials for applications in tactile sensors, wireless communicators, and multi-responsive soft actuators (adapted from ref. 38 and 42–45).

Heterogeneous blending of 2D materials within soft matter

Heterogeneous blending is a simple and straightforward approach to integrating a wide range of 2D materials into various soft matter for the fabrication of 2D material–soft matter composites. The heterogeneous mixtures of 2D materials and soft matter can be achieved by (i) mixing 2D material nanosheets with precursor solutions of elastomers or hydrogels followed by in situ crosslinking/curing or (ii) blending 2D material nanosheets into pre-synthesized/fabricated soft matter. The soft matter incorporated with 2D materials demonstrated enhanced mechanical, thermal, and electrical properties for soft actuators as well as provided built-in tactile sensing and remote control capabilities, showing their promising potential in developing multifunctional robotic materials.

Similar to other forms of nanomaterials (e.g., 0D nanoparticles, 1D carbon nanotubes)^46,47 that require additional functionalization and decoration for better integration within soft matter, necessary treatments, such as chemical functionalization^30,48 and polymer anchoring,^49–53 are normally required for 2D materials. For example, as emerging conductive 2D building block units, Ti₃C₂T_x MXenes are normally with the functional groups –OH, –F and –O, enabling their effective integration with various hydrogel-based soft matter.⁵⁴ Owing to the high electrical conductivity of MXenes (∼2000–6500 S cm⁻¹),⁵⁵ the MXene-integrated soft matter was found to be highly applicable in tactile sensing. Recently, Zhou et al. blended MXene nanosheets into a pre-fabricated, commercial PVA hydrogel to achieve a MXene–PVA hydrogel (M-hydrogel, Fig. 2A(i)).⁵⁴ The resulting M-hydrogel exhibited high strain sensing performance with a gauge factor of 25, which was higher than those of previously reported hydrogel-based strain sensors (Fig. 2A(ii)). On the other hand, the M-hydrogel could sense compressive strains and achieved a higher gauge factor of 80, due to the face-to-face interconnections under compression instead of the face-to-edge interconnections under tension. In addition, the M-hydrogel showed ultra-ductility (more than ∼3400%, Fig. 2A(iii)) and great adhesiveness, allowing M-hydrogel sensors to attach various surfaces and provide real-time sensing, including but not limited to facial expression recognition and vocal signal sensation (Fig. 2A(iv)).

Fig. 2 Heterogeneous blending of 2D materials within soft matter. The 2D material-blended soft matter demonstrated promising applications in tactile sensing (A and B), soft actuators (C–E) and light-driven robots (F–H). (A) (i) MXene-blended PVA hydrogel (M-hydrogel) showed high tactile sensitivity under both tension and compression loading. (ii) M-hydrogel showed great ductility, (iii) a high gauge factor of 25 for strain sensing and (iv) vocal signal sensation capability. Adapted from ref. 54. (B) (i) Graphene-putty composite (G-putty) exhibited high ductility and excellent tactile sensing performances for various signals, including (ii) finger wags and (iii) heartbeats. Adapted from ref. 56. (C) (i) OH–BNNS incorporated PNIPAM composite hydrogels demonstrated (ii) improved thermal conductivity and (iii) enhanced performance of thermally-triggered dye release. Adapted from ref. 57. (D) (i) GO integrated PNIPAM gradient hydrogels showed large curvature actuations in response to IR irradiation. (ii) The bending actuation behaviors were controllable by varying the applied electric field during the fabrication of GO/PNIPAM composites. Adapted from ref. 61. (E) (i) rGO-blended poly-(AMPS-co-AAm) hydrogels exhibited controllable curvature actuations in response to an applied voltage of 10 V. (ii) The rGO-blended hydrogels demonstrated improved mechanical integrity and were applied as (iii) “cantilevers” and (iv) “grippers”. Adapted from ref. 62. (F) (i–iii) Microstructural gradient ELP–rGO composite hydrogels enabled the fabrication of (iv) NIR light-driven micro-crawlers and (v) NIR laser-controlled hand-shaped actuators with reversible bending/unbending behaviors. Adapted from ref. 63. (G) Artificial micro-fish constructed with (i and ii) asymmetric rGO–PDMS/PDMS composites performed (iii) forward, backward, and turning motions upon the corresponding NIR light irradiation. Grid dimensions, 5 mm × 5 mm. Adapted from ref. 64. (H) (i) WS₂–PNIPAM composite hydrogel with (ii) precisely controlled actuations enabled the fabrication of (iii) a jellyfish-inspired micro-robot with (iv) swimming behaviors under periodic NIR irradiation. Grid dimensions, 5 mm × 5 mm. Adapted from ref. 66.

Besides hydrogels as introduced above, highly viscoelastic elastomers also served as host soft matter for integrating 2D material nanosheets. Recently, Boland et al. blended pristine graphene nanosheets into a slightly cross-linked polysilicone (Silly Putty) to obtain “G-putty” with unique electromechanical behaviors.⁵⁶ After the integration of pristine graphene, the G-putty still retained viscoelastic characteristics and showed high deformability (Fig. 2B(i)). During the deformations of the G-putty, the graphene conductive networks were broken and reformed reversibly in a time-dependent manner, and highly sensitive responses in the conductivity of the G-putty were observed. The G-putty was able to sense various mechanical triggers, including joint motions (Fig. 2B(ii)) and heartbeats (Fig. 2B(iii)). During the sensation of aortic pressure, the unprecedented sensitivity of the G-putty enabled the detection of characteristic double peaks and dicrotic notches with extremely fine resolution, from which the expected pulse (blood) pressure could be extracted after calibration. The high sensitivity was further highlighted by detecting the individual steps of a small spider (∼20 mg) walking over a cling film-coated G-putty sensor.

As a representative of insulative 2D materials, h-BN nanosheets (BNNS) are normally functionalized with the –OH and –NH₂ groups via multiple approaches including high temperature steam treatment, chemical oxidation, and mechanical ball milling.^30,57,58 Assisted by the hydrophilic functional groups, the functionalized BNNS could be well dispersed within precursor solutions of hydrogels followed by in situ polymerization or crosslinking, leading to uniform distribution of nanosheets within the hydrogels. Xiao et al. recently prepared –OH functionalized BNNS (OH–BNNS) via hot steam treatment at 850 °C, which were subsequently introduced into a precursor solution of poly(N-isopropylacrylamide) (PNIPAM) followed by in situ UV crosslinking. Attributed to the strong intermolecular hydrogen bonding between OH–BNNS and PNIPAM chains (Fig. 2C(i)), the mechanical properties of the OH–BNNS/PNIPAM composite hydrogels were significantly enhanced.⁵⁷ Meanwhile, the OH–BNNS with intrinsically high thermal conductivity (∼360 W m⁻¹ K⁻¹) could serve as efficient thermal carriers and led to improved heat transport properties of the composite hydrogels. As shown in Fig. 2C(ii), with only 0.07 wt% of OH–BNNS incorporated, the OH–BNNS/PNIPAM hydrogels showed 41% improvement in thermal conductivity as compared to bare PNIPAM hydrogels. As a result, the OH–BNNS/PNIPAM hydrogels exhibited more significant thermal actuating behaviors and faster thermally-triggered dye release (Fig. 2C(iii)).

GO nanosheets have also been considered as effective 2D nanofillers for hydrogel-based soft matter, because of their rich surface functional groups (e.g., –COOH, –OH) and high hydrophilicity.^59,60 Yang et al. developed GO/PNIPAM gradient hydrogels, where the distribution of GO nanosheets was controlled by electric fields prior to in situ UV crosslinking.⁶¹ The GO integrated hydrogels demonstrated enhanced light-to-heat conversion efficiency and contributed to improved photothermal actuations. The distribution gradient of GO within the PNIPAM hydrogels was ascribed to the gradient of electric fields, which led to uneven localized heating under IR radiation and thus unequal degrees of dehydration of the GO/PNIPAM hydrogels. The GO/PNIPAM gradient hydrogels demonstrated large actuation responses with directional bending (Fig. 2D(i)). Meanwhile, the actuation responses could be controllably enhanced by engineering the electric fields prior to the polymerization process: larger electric fields resulted in larger gradients of GO nanosheets within the composite hydrogels, leading to more uneven dehydration status and larger curvature changes (Fig. 2D(ii)).

GO nanosheets can be further reduced back to reduced GO (rGO) with superior electrical conductivity, enabling the production of electrically conductive rGO-blended hydrogels. Recently, Yang et al. developed a composite hydrogel of rGO/poly(2-acrylamido-2-methylpropanesulfonic acid-co-acrylamide) (poly-(AMPS-co-AAm)), which showed improved mechanical properties and electric current-induced thermal actuations (Fig. 2E).⁶² The well dispersed rGO nanosheets acted as efficient conductive nanofillers and generated significant osmotic pressure for the composite hydrogels, resulting in rapid and remarkable deswelling and bending behaviors induced by the electrical current. As shown in Fig. 2E(i), the speed of the electric current-induced actuations of the rGO/poly-(AMPS-co-AAm) hydrogels could be controllably accelerated as the loading of rGO nanosheets increased. The residual –O groups of the rGO nanosheets formed strong hydrogen bonding interactions with poly-(AMPS-co-AAm) chains, resulting in enhanced tensile and compressive strengths as well as better structural integrity upon large deformations (e.g., stretching and knotting, Fig. 2E(ii)). The rGO composite hydrogels further served as soft “cantilevers” for lifting a PDMS block with 0.2 g weight (Fig. 2E(iii)) and soft “grippers” for manipulating an object with 15 mm length (Fig. 2E(iv)).

The ability to remotely control the motions of soft robots enables their potential applications ranging from biomedical devices to manufacturing processes.¹⁷ Researchers have investigated various stimuli (e.g., light, magnetic fields, ultrasounds, electromagnetic waves) to remotely control the actuations of soft robots, which can be applied without physically connecting to the robotic bodies. In the following paragraphs, the advances of light-driven soft robots/actuators are summarized, demonstrating remotely controlled micro-robots or micro-motors with potential applications in cell/drug delivery, artificial muscles, etc. For instance, Lee et al. showed the noncovalent functionalization of rGO nanosheets using elastin-like polypeptides (ELPs),⁶³ after which the resulting ELP–rGO composite hydrogels were further engineered to obtain anisotropic porosity distribution along the composite thickness (Fig. 2F(i)). This anisotropic microstructure was induced by uneven water vapor exposure during the hydrogel crosslinking process. The top side of the ELP–rGO hydrogel exposed to rich water vapor showed higher porosity (Fig. 2F(ii)), while the bottom side with limited water exposure was relatively non-porous (Fig. 2F(iii)). As a result, the gradient ELP–rGO composite curled in response to cooling and heating, attributed to the faster swelling and deswelling behaviors of the top porous layer, respectively. Based on this mechanism, a light-driven micro-crawler was fabricated (Fig. 2F(iv)), which exhibited crawling behaviors consisting of curling/uncurling induced by the ON/OFF of near infrared (NIR) laser illumination. In the meantime, a hand-shaped ELP–rGO hydrogel actuator was designed (Fig. 2F(v)) to demonstrate its rapid, reversible, and tunable NIR-light-induced actuations, where the fingers showed bending and unbending in response to the varying location of the NIR laser spot.

In most cases, the actuations of hydrogel-based soft robots are restricted upon full immersion within aqueous solutions, where the unbalanced swelling behaviors are severely deteriorated. Therefore, other forms of light-driven robotic materials suitable for aqueous environments are demanded.¹⁷ To address this challenge, an alternative approach is to disperse rGO nanosheets into widely applicable elastomer matrices, enabling the design of (rGO-elastomer)/elastomer asymmetric structures for remotely controlled soft robots in a universal environment. Recently, Jiang et al. introduced rGO flakes into poly(dimethylsiloxane) (PDMS) (Fig. 2G(i)) to significantly alter its thermal (especially the coefficient of thermal expansion (CTE)) and mechanical properties.⁶⁴ Based on this, an asymmetric rGO–PDMS/PDMS robotic material was constructed (Fig. 2G(ii)), which could be actuated by NIR irradiation attributed to the mismatches in the CTEs and Young's moduli of the two layers. Furthermore, as shown in Fig. 2G(iii), an artificial micro-fish was demonstrated with a remotely controlled velocity and direction in both air and water environments upon NIR light illumination. Considering the controllable bending behaviors upon exposure to different incident directions of NIR light,⁶⁵ additional variations and adjustments could be made to this rGO–PDMS/PDMS robotic material to achieve more finely tunable actuations.

In addition to graphene and its derivatives, other 2D nanomaterials, such as TMDs, have been applied to fabricate light-actuated robotic materials for remotely controlled micro-robots. Recently, Zong et al. reported the excellent actuating behaviors of an ice-templated PNIPAM composite hydrogel containing alginate-exfoliated tungsten disulphide (WS₂) nanosheets (Fig. 2H(i)), which was actuated by the volumetric swelling and shrinkage of their cellular microstructures that mimicked plant cells.⁶⁶ This WS₂–PNIPAM hydrogel exhibited a fast actuation speed (up to 10° s⁻¹), high NIR light responsiveness, and a precisely controlled deformation direction upon varying the NIR radiation position (Fig. 2H(ii)), attributed to the nanosheet-reinforced pore walls and hierarchical water channels that facilitated efficient water translation. On top of this, a NIR light-driven jellyfish-inspired micro-swimmer was fabricated (Fig. 2H(iii)), which exhibited swimming behaviors under periodic NIR irradiation (Fig. 2H(iv)).

Bilayer integration of 2D materials onto soft matter

The aforementioned heterogeneous blending method has been demonstrated to be a straightforward strategy to confer the superior physicochemical properties of 2D materials to the fabricated 2D material–soft matter composites. However, it has to be admitted that, by heterogeneously blending 2D materials into soft matter, the physicochemical properties of 2D materials (e.g., thermal conductivity and electrical conductivity) were inevitably comprised, which may restricts the functionalities of resulting composites. In addition, with random distribution of 2D nanosheets, the heterogenous blending approach is not suitable for the applications that require high electrical conductivity and wide-spectrum chemical protection. Therefore, an alternative integration approach is highly desired to endow “hard” 2D materials with great stretchability/deformability while well preserving their intrinsic physiochemical properties.

An effective route has been demonstrated by integrating higher dimensional 2D material nanocoatings onto soft elastomers or hydrogels to achieve bilayer hybrid structures with high stretchability and well preserved intrinsic characteristics of 2D materials.^{38,39,67–69} The basic principle is to harness the surface instability during the contraction of a pre-stretched elastomer to programmably deform the top-layer 2D material nanocoating, achieving crumpled or wrinkled topographies. The accordion-like 2D material nanocoating is able to undergo reversible folding/unfolding to mitigate the in-plane strains under stretching, rendering excellent strain tolerance for reversible actuations. The bilayer configuration not only exhibits excellent mechanical stability and high conformability, but also preserves the physicochemical properties of 2D materials (e.g., barrier performance, electrical conductivity) under deformations, enabling their applications for stretchable protections and tactile sensing, which are both desired functions for soft robots.

In the section below, we will introduce a variety of 2D material/soft matter bilayer devices with large stretchability, which are expected to provide skin-mimicking capabilities (e.g., high stretchability, threat protection, tactile sensing, and conformal cell encapsulation) for soft robots. We further include several studies using 2D materials to fabricate flexible wireless communication and charging devices, which are important for the development of untethered robotic systems. Although limited stretchability was present in these wireless 2D material-based devices, we believe that, by following a similar strategy to fabricate textured 2D material–soft matter bilayer structures, wireless 2D material devices will be eventually adapted to soft robots and demonstrate deformation-insensitive performance.

The superior chemical and thermal stabilities of 2D materials (e.g., graphene, hBN) make them promising for their applications in protective barriers that can inhibit the permeation of undesired molecules^70,71 and prevent fire from the beginning.⁷² The broad-spectrum and durable protection enabled by 2D materials are highly desired for emerging soft robots, which can greatly expand their working environment to extreme conditions where accidental chemical spills or fire hazards are anticipated. For instance, GO films have shown high promise as protective nanocoatings with wide-spectrum chemical resistance,^70,73,74 resulting from the regularly stacked GO multilayers that are impermeable to diverse harsh chemicals.⁷⁵ However, since planar GO nanocoatings showed limited stretchability and tended to fracture even under minor in-plane strains, it has been challenging to apply conformal GO nanocoatings onto stretchable or deformable devices such as soft robots. Even though high stretchability is endowed by using the blending method, the formation of GO multilayers is not observed within the resulting GO–soft matter composites, which severely sacrifices their chemical protection performance.

To address this challenge, Chen et al. developed a bilayer hybrid structure consisting of a pre-crumpled GO nanocoating on top of a latex substrate, which was utilized as an ultra-stretchable molecular barrier (Fig. 3A).⁷⁶ The fabrication procedure started with transferring planar GO films onto pre-stretched latex substrates, where the strong adhesion between the GO and latex was facilitated by oxygen plasma pre-treatment. After the pre-stretched latex substrate was programmably contracted, the planar GO film was omnidirectionally compressed into a higher dimensional structure, and the as-obtained GO-latex bilayer device was able to sustain the stretching/relaxation fatigue tests under high areal strains up to 1500% for 500 cycles. The stretchable GO–latex devices were then verified to be high-performance molecule barriers against a variety of small molecule organic liquids even under extreme deformations. As a demonstration, a nitrile glove protected with textured GO nanocoatings was able to endure the permeation of dichloromethane (DCM) for 30 minutes submersion. The mechanically stable and stretchable GO–elastomer devices are expected to be useful for serving as protective skins for soft robots working in extreme circumstances where exposure to harsh chemicals is anticipated.

Fig. 3 Bilayer integration of 2D materials onto soft matter. The bilayer 2D material–soft matter structures enabled the applications of stretchable protection (A, B and C), stretchable tactile sensation (D, E and F), bio-encapsulation and sensing (G, H and I), as well as wireless communication (J, K and L). (A) Textured GO–latex bilayer structure for ultra-stretchable molecular barriers against various organic solvents. Adapted from ref. 76. (B) Textured MMT–elastomer bilayer structures for fire-retardant barriers to protect a soft robotic gripper working in a fire scene. Adapted from ref. 68. (C) Monolayer graphene was conformally coated onto a commercial contact lens and demonstrated effective EMI shielding and dehydration protection performance. Adapted from ref. 77. (D) A stretchable pressure sensor was fabricated based on hierarchically textured rGO–elastomer electrodes, showing strain-insensitive pressure sensing capability that was used for smart “collision aware” surgical robots. Adapted from ref. 18. (E) Layer-by-layer (MXene–PDAC)_n composites were stretchable and demonstrated human motion detection performance. Adapted from ref. 78. (F) Deformable WS₂–PDMS sensor showed stable humidity detection performance upon bending. Adapted from ref. 79. (G) Flower shaped FLG/SU-8 bilayer device showed self-folding behavior upon solvent changing. Scale bar, 200 μm. Adapted from ref. 82. (H) Ultrathin thermally responsive functionalized graphene flower reversibly folded and unfolded upon temperature change, enabling conformal wrapping of breast cancer cells. Scale bar, 100 μm. Adapted from ref. 83. (I) Further modifying the self-folded graphene system with Ag NCs led to an ultrathin, flexible, skin-like biosensing platform, enabling significantly enhanced Raman response and 3D label-free spatially resolved surface analysis. Adapted from ref. 81. (J) Flexible UHF RFID antenna based on graphene showed consistent performance under deformations. Adapted from ref. 84. (K) Flexible MXene antenna with a controllable reflection coefficient by varying the thickness of MXene nanocoatings. Adapted from ref. 85. (L) Atomically thin and flexible MoS₂ rectenna exhibited high power efficiency conversion at the Wi-Fi band, which was comparable with the state-of-the-art results. Scale bar, 100 μm. Adapted from ref. 86.

In addition to stretchable molecular barriers, 2D material nanocoatings can act as efficient fire-retardant layers to be integrated onto soft robots, allowing the robots to execute tasks under high-temperature conditions or even fire scenarios that are not directly accessible for humans. As one of the frequently-used 2D materials to enhance the flame retardancy of targeted objects, montmorillonite (MMT) has been intensively investigated. Similar to all other 2D material thin films, MMT nanocoatings rather undergo mechanical fracture upon stretching, which severely restricts their further applications in mechanically dynamic protection. To tackle this challenge, Chang et al. integrated crumpled MMT nanocoatings onto elastomer substrates, and the as-fabricated bilayer MMT–elastomer devices showed combinative features of a high strain tolerance of up to 225% areal strain and effective flame retardancy in the meantime.⁶⁸ In contrast to the bare elastomer that burned within 5 seconds under direct flame contact, the MMT-protected elastomer showed largely improved flame retardancy that endured the fire for over 60 seconds even under a uniaxial strain of 100%. The MMT–elastomer bilayer composite was further utilized as the flame-retardant skin for soft robots. As shown in Fig. 3B, with the protection of crumpled MMT nanocoatings, a pneumatic soft gripper was actuated upon direct contact with fire and executed the rescuing task without being ignited, showing promise for application in fire-retardant skin for soft robots working under extreme conditions.

Another ubiquitous physical threat that may affect normal functions of soft robots, especially those embedded with complex electronic systems, is electromagnetic interference (EMI). In this case, graphene can be of great value for EMI shielding barriers, due to its outstanding electrical properties. Recently, Lee et al. transferred a monolayer of chemical vapor deposition (CVD)-grown graphene onto a commercial contact lens (Fig. 3C).⁷⁷ Upon EM wave shielding tests on egg whites inside a microwave oven, it was shown that the EM energy was absorbed by graphene and further dissipated in the form of thermal radiation to minimize the possible EM damage. The IR camera image demonstrated that the temperature of the graphene-coated contact lens increased greatly as compared to the constant temperature of the commercial lens in the microwave oven, indicating that the EM energy was efficiently absorbed by the graphene coating and further dissipated as heat. Meanwhile, the conformally-coated graphene served as the barrier layer for moisture, which prevented the dehydration of the contact lens during long practical use. The conformal graphene layer exhibited high EMI shielding performance and dehydration protection as well as excellent biocompatibility, showing its great potential in serving as a protective layer for soft robots particularly designed for bio-related systems.

Besides stretchable and conformal barrier characteristics, the bilayer configuration of 2D material–soft matter devices enables the preservation of the high electrical conductivity of 2D materials under large strains, holding promise for their application in stretchable tactile sensing. As Chang et al. recently reported, stretchable rGO electrodes with hierarchical Shar-Pei-like topographies were fabricated by a two-step deformation process (Fig. 3D).¹⁸ The specially-designed rGO electrode with a crumpled–wrinkled topography exhibited a strain-invariant resistance profile (with variations less than 3%) under reversible loadings under uniaxial strains up to 100%. When two stretchable rGO electrodes were clamped face-to-face (Fig. 3D(i)), the fabricated pressure sensor was able to differentiate “soft” and “hard” presses with high sensitivities of 1.37, 1.30, and 0.98 kPa⁻¹ under 0%, 30%, and 50% uniaxial strains, respectively. The stretchable pressure sensor with insensitive strain interference demonstrated a promising application in transoral robotic surgery (TORS), providing tactile “awareness” to unpredictable accidental collisions during the robotic surgery (Fig. 3D(ii)).

Another similar strategy to fabricate highly deformable 2D material–soft matter composites with preserved functionalities of 2D materials is through the construction of a layer-by-layer (LBL) structure by alternatively depositing 2D material nanosheets and polymers. An et al. recently reported LBL-assembled nanocoatings by sequentially adsorbing negatively-charged MXene nanosheets and positively-charged poly(diallyldimethylammonium chloride) (PDAC), as shown in Fig. 3E(i).⁷⁸ The (MXene–PDAC)_n nanocoatings were demonstrated to be stretchable, bendable, and electrically conductive, enabling conformal adhesion onto various surfaces ranging from silicon, glass, nylon fiber, and PDMS to flexible polymer sheets. Moreover, these (MXene–PDAC)_n nanocoatings well preserved the electrical conductivity (∼2000 S m⁻¹) of the MXenes and demonstrated large deformability. In particular, these LBL-assembled nanocoatings demonstrated sensitive and recoverable resistance responses to cyclic stretching (up to 40% strain) and bending (up to 2.5 mm radius, Fig. 3E(ii) and (iii)). These high-sensitivity (MXene–PDAC)_n sensors were able to detect topographical differences and various human motions upon appropriate implementation onto fingers, indicating their high potential in serving as E-skins for soft robots.

In addition to stretchable tactile sensing, 2D materials can be applied for fabricating chemical sensors such as water vapor (i.e., humidity) and gas molecule detectors, which are highly demanded for the development of intelligent soft robots perceiving environmental changes. As soft robotic backbones are not suitable for implementing conventional rigid transistors, the development of stretchable chemical sensors with strain-invariant performance is in urgent demand. As Guo et al. recently reported, a stretchable and transparent WS₂-based humidity sensor was fabricated (Fig. 3F).⁷⁹ The as-fabricated WS₂ sensor exhibited high humidity sensing performance over a wide relative humidity range up to 90%, and the detection sensitivity of the sensor was free of any impact from various mechanical deformations such as bending, compression, and stretching. Furthermore, the WS₂ humidity sensor could be well laminated onto human skin, showing stable moisture sensing and real-time breath monitoring under various human body motions, shedding light on the development of energy-efficient and deformable environmental sensors for soft robots.

Besides the outstanding multifunctionalities under various mechanical loadings, the 2D material–soft matter bilayer structures can be applied for the fabrication of micro- or nano-scale robots as well. By applying single- or few-layer 2D materials on thin stimuli-responsive substrates, thin bilayer devices were obtained, which could exhibit superior conformability over micro-scale and irregular shaped objects (e.g., pollen grains, beads, and live cells) upon appropriate design and fabrication.^80,81 Recently, bilayer integration of continuous monolayer graphene with stimuli-responsive polymer substrates has been reported to be an effective approach to fabricate reconfigurable structures with desired patterns and functionalities. Deng et al. transferred CVD grown few layer graphene (FLG) onto a gradient-cross-linked epoxy (SU-8) film (Fig. 3G(i)),^81,82 achieving a bilayer FLG/SU-8 device that could fold and unfold reversibly upon changing the solvent in its surroundings (Fig. 3G(ii) and (iii)). Further incorporation of sensing electrodes into this structure could augment its function as 3D bio-/chemical sensors.

The thickness of the FLG/SU-8 bilayer was reported to be ∼10–100 μm, which was relatively thick to fully utilize the superior flexibility and conformability of 2D materials. Therefore, the fabrication of ultrathin bilayer structures with thicknesses of a few nanometres or even less becomes essential. To tackle this challenge, Xu et al. demonstrated thermally triggered folding and unfolding of monolayer graphene by functionalizing its surface with polydopamine (PD) and PNIPAM brushes, and the functionalized monolayer graphene was abbreviated as G-PD–PNIPAM.⁸³ The micropatterned G-PD–PNIPAM film was as thin as 5 nm, which could self-fold into 3D structures following pre-designed patterns under heating (Fig. 3H(i)–(iii)). Moreover, this self-folding behavior was reversible and tunable by simply varying the temperature within a mild temperature range that was compatible with cell culture and many other biological processes. As shown in Fig. 3H(iv), the graphene flower was able to conformally encapsulate a cancer cell. These features enable the promising potential of this reconfigurable 3D graphene robotic backbone in the applications of bioelectronics, biosensors, and bio-delivery systems.

On top of the highly conformable self-folding behaviors, further modification has been introduced to augment additional 3D biosensing capabilities.⁸¹ Very recently, silver nanocubes (Ag NCs) were further decorated onto functionalized monolayer graphene to achieve G-PNIPAM–G-Ag films. The resulting ultrathin hybrid film was able to conformally encapsulate soft 3D biological samples with irregular shapes (e.g., breast cancer cells, as shown in Fig. 3I(i) and (ii)). Attributing to the decoration of Ag NCs, the ultrathin and flexible skin showed surface-enhanced Raman spectroscopy (SERS) performance with an enhancement factor of up to 110, in contrast to the planar G-PNIPAM–G-Ag film and bare quartz substrate that exhibited limited Raman signal (Fig. 3I(iii)). Furthermore, the strong Raman signals from the cell membrane facilitated 3D label-free spatially resolved surface analysis (Fig. 3I(iv)), which was not achievable for conventional planar rigid SERS substrates due to their relatively limited contact area.

To enable the remote operations of soft robots in in vivo environments (e.g., human body), conventional approaches are to use magnetic fields or ultrasounds to drive the movements of robots, which could be invasive to some extent and troublesome in the meantime with additional magnetic functionalization. Alternatively, 2D materials with intriguing electromagnetic properties have been reported with promising wireless sensing and communicating capabilities, which make 2D materials attractive for the fabrication of soft robots that are capable of being remotely controlled or powered by electromagnetic waves. In addition, with the rapid development of the Internet of Things, the capabilities of wireless communication/charging/actuation are becoming highly demanded for next-generation intelligent soft robots in the upcoming wireless Internet era. As such, we would like to introduce recent studies using 2D materials for fabricating flexible wireless devices. To further apply 2D material-based wireless devices onto untethered soft robots, we anticipate that additional stretchability is required for wireless devices, which may be achieved by the bilayer configuration with the highly preserved physicochemical properties of 2D materials and the high elasticity of soft matter.

Recently, Pan et al. produced highly conductive graphene ink that could be directly printed onto flexible substrates such as polymer textiles and papers (Fig. 3J(i)).⁸⁴ The printed graphene devices demonstrated high electrical conductivity (7.13 × 10⁴ S m⁻¹) and enabled the fabrication of a wireless connectivity antenna that operated over a wide frequency range from MHz to tens of GHz. The flexible antenna with ultra-high frequency radio-frequency identification (UHF RFID) exhibited a long communication range of >9 m and covered an overall UHF RFID band (860–960 MHz) (Fig. 3J(ii)). A typical dipole pattern was observed for the radiation of the as-fabricated antenna with a minimum reading range at 90° and 270° and a maximum reading range at 0° and 360° (Fig. 3J(iii)).

As one of the water-dispersible and electrically conductive 2D materials, MXenes open new pathways for the fabrication of wearable, flexible, and portable RF communication devices. Recently, lightweight and flexible antennas based on 2D metallic MXene nanosheets were achieved by Sarycheva et al. using scalable spray coating.⁸⁵ With the thickness of the MXene antennas varying from ∼100 nm to 8 μm, the reflection coefficient changed from −10 to −65 dB and was highly coherent with the simulation results at varying frequencies (Fig. 3K). The performances of MXene antennas with much lower thicknesses were higher than those of the best-known nanomaterial antennas with comparable thicknesses (e.g., silver paste or graphite ink), enabling the manufacture of transparent and ultrathin wireless devices, as well as their promising applications in the Internet of Things and next-generation untethered soft robots.

With Wi-Fi systems penetrating into almost every scenario of our daily life, the accompanying EM radiation becomes an ideal source for powering future wireless electronics and untethered soft robots. Recently, an atomically thin and flexible radiofrequency harvester (rectenna) based on a MoS₂ semiconducting-metallic-phase heterojunction has been developed (Fig. 3L(i) and (ii)).⁸⁶ Compared to state-of-the-art flexible semiconductor-based rectennas that cannot work at high frequencies for Wi-Fi communications, the MoS₂-based rectenna showed a high cutoff frequency of 10 GHz, covering not only unlicensed medical, scientific, and industrial radio bands (including the Wi-Fi channels), but also the X-band (8 to 12 GHz). The RF-d.c. power conversion efficiency of this as-fabricated rectenna in the Wi-Fi band (2.4 and 5.9 GHz) was further calculated (Fig. 3L(iii)). At 2.4 GHz, the power efficiency was extracted to be 40.1% when a power of −0.7 dBm was inputted, which was comparable with that of the state-of-the-art rigid-diode technology. After further integration with a flexible Wi-Fi band antenna, a fully self-powered antenna was achieved. Although these flexible antennas based on 2D material-flexible substrate bilayer structures have been demonstrated, their stretchability is still a bottleneck that severely restricts their further application in mechanically dynamic soft robots, where various tensile loadings are anticipated. To address this, more research efforts await to fully take advantage of the high conductivity of 2D materials and the elasticity of soft matter, enabling their practical applications in stretchable and deformable antennas (rectennas) for soft robots.

Post-stabilization of 2D material (or 2D material-templated) architectures with soft matter

Another strategy to produce multifunctional robotic materials is through the infiltration/encapsulation of thin soft matter into/onto pre-assembled 2D material architectures. This post-stabilization approach has been considered as an effective method to improve the mechanical stability of 2D material assemblies while also mostly preserving their intrinsic characteristics. For instance, 3D graphene aerogels (GA) or graphene foams (GF) demonstrate well-controlled structures and high electrical conductivities, yet the highly porous 3D graphene structures are fragile and lack structural deformability, shape recoverability, and flexibility, which severely restrict their applications in mechanically dynamic devices. To address this issue, Li et al. recently introduced a thin PDMS layer with controllable thickness onto bare GF synthesized by a CVD method (Fig. 4A(i) and (ii)).⁸⁷ As compared to the bare GF that plastically deformed upon compression, the stabilized GF@PDMS composite exhibited much higher mechanical deformability, better structural stability, and larger shape recoverability upon exposure to various mechanical loadings, such as compression, twisting, and bending (Fig. 4A(iii)). After the stabilization with the thin PDMS coating, the EMI shielding performance was largely preserved for GF@PDMS (up to 36.1 dB) over a wide frequency range of 8.2–18.0 GHz, which was even free of the influence from the various mechanical loadings. The deformation-insensitive EMI protection is highly demanded for functional soft robots embedded with complex electronic systems.

Fig. 4 Post-stabilization of 2D materials (or 2D material-templated) architectures with soft matter. The pre-assembled 2D materials (or 2D material-templated) architectures stabilized with soft matter enabled their applications in deformable protection (A and B), tactile sensation (C and D), and multifunctional robotic backbones (E and F). (A) Graphene foams (GF) stabilized with thin PDMS coatings showed greatly enhanced mechanical properties and well-maintained EMI shielding performance. Adapted from ref. 87. (B) 3D-BNNS/PDMS architectures were highly deformable and showed excellent thermal regulation capability. Adapted from ref. 88. (C) Selective gluing of graphene aerogel (GA) with PDMS resulted in a deformable conductive composite with durable pressure sensing performance. Adapted from ref. 89. (D) PDMS-stabilized GA performed strain-dependent electromechanical response and reliable temperature stability. Adapted from ref. 90. (E) GO-enabled templating method led to the fabrication of MO origami robots, which showed various robotic actuations after elastomer stabilization. The legendary phoenix-fire-reborn concept was realized, where a paper origami robot sacrificed itself in fire and transformed into a downsized Al₂O₃ robot. (F) GO-enabled templating method produced noble metal-based origami robots that exhibited built-in strain sensing and wireless communication capabilities. Adapted from ref. 25.

Very recently, Hou et al. prepared a mesoporous BNNS network with unidirectional micro-channels via direct freezing followed by carbonization welding (Fig. 4B(i) and (ii)).⁸⁸ The mesoporous BNNS network was further stabilized with a thin PDMS coating, and the interconnected BNNS structures were well preserved. The resulting 3D-BNNS/PDMS structures showed a 3900% improvement in thermal conductivity in comparison with bare PDMS (Fig. 4B(iii)) and were able to sustain cyclic twisting and bending deformations. The IR thermal images during the cooling process indicated that the heat dissipation rate of the composite was faster than the random-BNNS/PDMS composite and bare PDMS. This deformable heat dissipative 3D-BNNS/PDMS composite showed promising potential in the heat regulation for soft robotics, which could be of great significance considering the aggravating heat accumulation caused by the increasing level of system integration.

Besides fully immersing 2D material assemblies into elastomers that may risk sacrificing the intrinsic characteristics of 2D materials (e.g., thermal conductivity, electrical conductivity), an alternative method is to provide only necessary connections between 2D nanosheets and selectively glue the intersheet joints. Being inspired by the soft cartilage connecting hard bones, Hong et al. demonstrated an ultra-durable and super-elastic 3D GA@PDMS composite (Fig. 4C(i)).⁸⁹ The selective inclusion of viscoelastic PDMS into the junctions of the graphene nanosheets was facilitated by the capillary forces of diluted PDMS solution. The resulting 3D GA@PDMS architectures exhibited well-preserved high porosity as well as excellent conductivity. These features provided effective stress-transfer pathways that enable excellent shape recoverability of GA@PDMS under repeated compressions with strains of up to 90% and reliable strain sensing capability with very short response times (Fig. 4C(ii) and (iii)). The cartilage-inspired GA@PDMS architectures opened new possibilities for fabricating soft robotic materials with high structural stability, shape compliancy, and built-in strain sensing capability.

Similarly, Zhang et al. produced a PDMS-stabilized 3D GA framework via ice-bath-assisted infiltration and vacuum curing processes (Fig. 4D(i)).⁹⁰ With the well-maintained interconnections between graphene nanosheets, the GA–PDMS composites (GAPC) showed large shape deformability (with tensile and compressive strains of up to 90% and 80%, respectively), rapid electric Joule heating performance ((dT/dt)_max > 3 °C s⁻¹ under a heating power of 12 W cm⁻³), high electrical and thermal conductivities (1 S cm⁻¹ and 0.68 W m⁻¹ K⁻¹, respectively), and stable piezo-resistance performance. As a demonstration, the GAPC exhibited typical non-linear resistivity response under cyclic sine-wave dynamic compressions (0.1 Hz, Fig. 4D(ii)) and gradually stabilized electromechanical responses upon cyclic bending–flatting deformations. Further electrical characterization studies at various temperatures showed that the electrical conductivity of the GAPC was temperature-insensitive in comparison with bare GA (Fig. 4D(iii)).

Besides direct use of 2D materials for the fabrication of multifunctional robotic materials, 2D materials with high surface reactivity can be adopted as synthetic templates to produce other nanomaterials with new functionalities. In particular, attributed to their rich functional groups, ease of aqueous processing, and tunable interlayer spacing, multilayered GO films have been demonstrated as sacrificial templates for the nucleation and growth of various nanomaterials, especially metal oxides (MOs) and noble metals.⁹¹ Upon the intercalation of metal ions into GO multilayers, the interlayer galleries guided the conversion of metal ion precursors into lamellar MO structures during the high-temperature calcination process. The GO-enabled templating synthesis has been reported regarding the growth of various MO nanocrystals (e.g., YBa₂Cu₃O_7−δ (Y123),⁹² TiO₂, Fe₂O₃, and ZnO⁹³) with complicated macro-, micro-, and meso-scale structures replicated from their GO templates, such as freestanding strands,¹² wrinkled/crumpled textures,¹⁷ and planar multilayers.²⁰ The effective structural reproducibility of this GO-enabled synthesis makes it a promising approach towards the fabrication of multifunctional MO- and noble metal-based backbones for soft robots.

Very recently, Yang et al. adopted the GO-enabled templating method to transform complicated paper origamis into MO origami replicas with high structural reproducibility.²⁵ Followed by the infiltration of thin elastomer coatings, the resulting MO–elastomer origamis were stabilized and enabled the fabrication of reconfigurable MO robots with unconventional characteristics. The synthetic route of MO robotic backbones is schematically illustrated in Fig. 4E(i), which consists of four steps: (i) deposition of GO nanosheets onto cellulose paper origamis, (ii) intercalation of hydrated metal ions into GO–cellulose templates, (iii) high-temperature calcination for removing carbonaceous templates, and (iv) stabilization of MO replicas with elastomers. Various 3D origami structures in Al₂O₃, including auxetic hexagonal honeycomb and bellows tube structures, were successfully synthesized by this GO-enabled templating approach (Fig. 4E(ii)). After the infiltration of thin elastomers into Al₂O₃ origami structures, the Al₂O₃–elastomer origamis not only well replicated the folding patterns of their paper origami templates but also exhibited high reconfigurability and shape deformability (180° bending, 180° twisting, 60% stretching) (Fig. 4E(iii)). The MO–elastomer origamis can be considered as reconfigurable metamaterials and be further equipped with various actuation systems (e.g., pneumatics, shape-memory alloys and magnetic fields) to fabricate soft robots. Soft robots with MO backbones (abbreviated as MO robots) exhibited lower weight, higher energy efficiency, and better compliancy than conventional paper origami robots. As shown in Fig. 4E(iv), the legendary phoenix-fire-reborn concept was realized, where a paper origami robot sacrificed itself in fire and transformed into a downsized Al₂O₃ robot. The resulting Al₂O₃ robot then crawled through a narrow tunnel where the original paper origami robot cannot pass.

This GO-enabled templating method can be further applied to produce noble metal robotic backbones for the fabrication of multifunctional soft robots with built-in strain sensing and wireless communication capabilities (Fig. 4F). By following similar GO-enabled synthesis and PDMS stabilization processes, various metallic robotic backbones (e.g., Ag, Au, Pt) were produced with replicated origami patterns, high electrical conductivity, and large reconfigurability. The representative Pt-elastomer backbones enabled the fabrication of soft robots with diverse and unique functionalities, such as built-in strain sensing, on-demand resistive Joule heating, and deformable wireless communication (Fig. 4F(i)). As shown in Fig. 4F(ii), the bellows-type Pt robot showed stable strain sensing performance and reliable durability during the cyclic stretching/relaxation testing (up to 60% strain, 1.2 Hz). The built-in strain sensing capabilities of the reconfigurable Pt backbones allowed users to monitor and record robotic actuations in real time. In addition, the Pt robotic backbones could act as deformable antennas and imparted wireless communication capabilities to Pt robots. Fig. 4F(iii) shows that the pulse signals at different frequencies (sent by a sender Pt robot) were well received by a receiver Pt robot that was 1.2 m away, and no frequency deviation between the sent and received signals was observed.

We anticipate that soft origami robots with MO and noble metal backbones are competitive candidates for a wide range of applications, such as robots in harsh environments (e.g., chemical spills, high temperature/fire), artificial muscles, and humanoid robotic arms. More fascinating functionalities could be incorporated via the intercalation of various metal ions during the GO-enabled synthesis, which is expected to further expand the material library for fabricating multifunctional soft robots. For example, metallic backbones can be further functionalized with electrochemically active materials, allowing fabrication of energy-storage devices based on robotic backbones themselves.

Discussion

Various approaches have been developed for the integration of 2D materials with soft matter towards their potential applications in multifunctional robotic materials, which were defined as mechanically soft and reconfigurable materials with built-in sensing, actuating, protection, and wireless communication capabilities. As summarized in Table 1, these diverse functionalities are not achievable for conventional robotic materials, such as soft matter (e.g., hydrogels,^30,94–96 elastomers^16,95,97), shape memory alloys (SMAs),^98,99 and shape morphing polymers (SMPs),^98,100 most of which only perform actuation function. In addition, these 2D material–soft matter based soft robotic materials possess much larger maximum strain (ε_max, up to 3.0)^101,102 and widely tunable mechanical stiffness (Young's moduli, ∼10⁴–10¹² Pa),¹⁰³ which make them competitive candidates for fabricating soft robots with superior flexibility and compliance upon request. Furthermore, their broad density range (∼0.003–1.1 g cm⁻³)¹⁰⁴ and variable dimensions (lower than the nanoscale) enable the construction of robotic systems with sizes lower than 1.5 × 10⁻⁵ m,⁸³ which shows high promising potential in untethered miniature soft robots for biomedical applications.

Table 1 Comparison between 2D material–soft matter, soft matter (e.g., hydrogels, elastomers), and representative conventional robotic materials (e.g., shape memory alloys (SMAs), shape morphing polymers (SMPs)). ε_max and mechanical stiffness indicate maximum strain and Young's modulus, respectively

Robotic materials	ε max	Mechanical stiffness (Pa)	Backbone density (g cm^−3)	Smallest robot size (m)	Functionality
SMAs^98,99	0.07	∼10^10	∼6.0–8.0	4.0 × 10^−4	• Actuation
SMPs^98,100	1.0	∼10^9	∼0.9–1.2	1.0 × 10^−2	• Actuation
Hydrogels^30,94–96	0.9	∼10^4–10^7	∼1.0–1.1	1.2 × 10^−3	• Soft actuation
					• Chemical sensing
Elastomers^16,95,97	3.0	∼10^4–10^6	∼0.965–1.07	3.7 × 10^−3	• Soft actuation
2D material–soft matter materials^83,101–104	3.0	∼10^4–10^12	∼0.003–1.1	1.5 × 10^−5	• Soft actuation
					• Stretchable tactile/chemical sensing
					• Stretchable protection
					• Wireless communication

---

In terms of fabrication methodologies, three major approaches including heterogenous blending, bilayer integration, and post-stabilization have been reviewed in detail as above. Each approach exhibits its own technological advantages and drawbacks, as summarized in Table 2. Heterogeneous blending is considered to be the most straightforward, simple, feasible, and widely applicable method for integrating a wide range of 2D materials into soft matter, leading to their enormous possibilities in developing diverse robotic materials as demanded. This method imparts the excellent mechanical properties of soft matter and various functionalities of 2D materials to the as-fabricated 2D material–soft matter composites. However, the heterogenous blending approach is a relatively random process with imprecise control over the distribution of 2D materials within soft matter, the interconnections of 2D materials, and the interactions between 2D materials and soft matter. The lack of abilities to control the organization of 2D materials within soft matter at the nanoscale could affect the performance of potential robotic applications requiring high molecular rejection, high electrical conductivity, and accurate chemical sensing. Another concern that should be noted is that the blending method compromises the intrinsic characteristics of both 2D materials and soft matter; that is, the 2D material–soft matter composites exhibit combined but possibly worse performance in comparison with the intrinsic functionalities and mechanical properties of pure 2D materials and soft matter, respectively, which is an inevitable trade-off.

Table 2 Comparison between different approaches to fabricating 2D material–soft matter robotic materials with multifunctionality

Approaches
Advantages	• Simple/straightforward fabrication	• Fully exposed surfaces of 2D materials	• Well preservation of specifically designed 2D material architectures
	• Wide applicability	• Precise control of 2D materials’ morphologies and thicknesses
Drawbacks	• Limited control over distribution of 2D materials in soft matter	• Troublesome 2D material pre-texturing	• Complicated pre-fabrication of 2D material assemblies
		• Tricky interfacial engineering	• Non-controllable soft matter infiltration
		• Limited material candidates
Features	• Combined yet compromised physicochemical properties	• Largely preserved 2D material properties	• Largely preserved 2D material properties
		• Slightly increased mechanical stiffness	• Limited mechanical stretchability
Applications	• Soft functional actuators	• Stretchable protection	• Deformable physical protection
	• Tactile sensors	• Stretchable tactile/chemical sensors	• Durable mechanical sensation
		• Wireless communication	• Multifunctional robotic backbones

---

Unlike the approach of heterogeneous blending, the bilayer integration method produces a hybrid structure with a minimal compromise of the intrinsic properties of both 2D materials and soft matter. The fully exposed surfaces of 2D materials allow the resulting bilayer structure to take full advantage of the unique properties of the 2D materials, enabling sophisticated applications in chemical/gas sensing, wireless communications, energy storage/harvesting, and chemical/fire barriers. Meanwhile, the functionalities of 2D material–soft matter devices can be precisely tuned by thickness control,^105,106 topography engineering,^18,40,41 and even heterostructure fabrication of top-layer 2D materials.^29,107 Despite these merits, the fabrication of this bilayer configuration is rather complicated, which involves the design of 2D material textures and interfacial engineering between 2D material nanocoatings and soft matter. To be specific, the combinations of 2D materials and soft matter should be carefully selected to enable strong adhesion and thus synchronized mechanical actuations, narrowing down the selection of 2D material candidates to a small library. In addition, the topographies of top-layer 2D material nanocoatings need to be specially designed to ensure their structural integrity under reversible mechanical loadings. Furthermore, substantial efforts should be devoted to engineering the interfacial interactions between 2D materials and soft matter, such as hydrogen bonding,^18,108 van der Waals interactions,^107,109 and electrostatic interactions,^78,110 which are crucial for providing sufficient mechanical stability demanded by various robotic applications. The bilayer 2D material–soft matter robotic materials are more suitable for serving as the E-skin of soft robots that requires sensing, actuation, and barrier functionalities, rather than as the main robotic backbone or body due to their relatively thin thicknesses.

The post-stabilization method involves the infiltration of soft matter into pre-assembled 2D material frameworks, which thus enhances the mechanical deformability and still largely preserves the interconnected networks of 2D materials in the meantime. Owing to the well-maintained frameworks, the pre-designed characteristics of 2D material foams or aerogels, such as electrical conductivity, specific surface area, and ordered structures, can be mostly preserved. We consider the post-stabilization approach as a compromise of the aforementioned two methods. The resulting soft matter-stabilized 2D material composites exhibit limited stretchability, and most of them can only endure bending and reversible compression loadings. Meanwhile, the pre-fabrication of 2D material frameworks (e.g., foams, sponges, aerogels) could be troublesome to some extents, where methods including CVD, freeze-drying, or hydrothermal treatments may be involved. It should also be noted that the infiltration of soft matter is not fully controllable, which may lead to the random diffusion and non-homogeneous distribution of soft matter throughout the entire 2D material architectures, restricting their further applications in robotic materials where high structural uniformity is demanded.

Outlook

Considering the current trend of achieving lightweight, durable, and eventually untethered, self-powered soft robots with high levels of system integration,^3,45,111 enormous research efforts have been devoted to the development of multifunctional robotic materials. The advantage of using these multifunctional robotic materials is to augment the functions of soft robots without integrating external electronics or exoskeletons, which does not increase the overall weight of the robots and could simplify the system design in the meantime.

The future development of multifunctional robotic materials could be elaborated across multiple length scales. At the nano-/micro-scale, more work could be done to synthesize new 2D materials, hydrogels, and elastomers to expand the candidate library for the fabrication of functional robotic bodies/skin by following the aforementioned strategies. It is desired to synthesize new 2D materials (e.g., boron nanosheets (borophene),^112,113 graphdiyne¹¹⁴) with intriguing functionalities and introduce them into robotic materials to explore more possibilities. Besides, an alternative way would be to introduce functional groups^115–117 into or anchor polymer chains^118,119 to existing 2D material candidates, by which the physicochemical properties of 2D building block units can be systematically tuned. Another factor that should be considered for 2D materials is their lateral size, which has been reported to affect their reinforcing effects greatly in performances, such as optical,^120,121 thermal,^122,123 and mechanical properties.^124,125 As for soft matter, appropriate treatments or modifications can be applied to functional polymers (e.g., conjugated polymers),¹²⁶ shape memory polymers,¹²⁷ or elastomers to adjust their intrinsic characteristics for better compatibility with 2D materials. Furthermore, functional polymers in a 2D format are emerging candidates that may inspire a new category of multifunctional soft materials for soft robotics.^128–130

At the meso-scale, especially in the heterogeneous blending approach, research gaps still remain in achieving the controllable distribution of 2D materials within soft matter. The organization of 2D nanosheets has been reported to significantly affect the mechanical strength,^131,132 thermal conductivity,^133,134 and optical properties,^135,136 as well as provide augmented functionalities (e.g., chemical/tactile sensing,^137–139 actuation^36,140,141) of the resulting robotic materials. Meanwhile, direct additive manufacturing, such as 3D printing, inkjet printing and fused deposition modelling,¹ could be adopted to obtain structured or patterned 2D material–soft matter composites, which could facilitate the fabrication of soft robots with finely designed mesoscale architectures and thus wide adaptation to various actuation systems.

From our perspective, one of the ultimate goals in soft robotics fields at the macro-scale is to achieve a crew of untethered, lightweight, and self-powered soft robots that can wirelessly communicate and collaborate with others to execute sophisticated tasks. The collaborative crew of soft robots would promise a wide range of technologies from autonomous field robots to wireless biomedical robots, which are of great significance to the upcoming wireless Internet of Things era.^4,45 Therefore, research efforts should be particularly made towards imparting wireless functionalities to soft robots, such as wireless communication^85,142 and wireless/Wi-Fi charging,^86,143 considering that current wireless devices still do not possess sufficient deformability/stretchability for soft robotic applications. In addition, the biocompatibility of robotic materials is a crucial factor to be taken into account for the development of small-sized soft robots that are very useful in biomedical applications, such as the delivery of imaging agents, genes or drugs under wireless control.¹⁷

Conflicts of interest

There is no conflict of interest to declare.

Acknowledgements

The authors acknowledge the financial support provided by the Faculty Research Committee (FRC) Start-Up Grant of the National University of Singapore R-279-000-515-133, the Ministry of Education (MOE) Academic Research Fund (AcRF) R-279-000-532-114, R-279-000-551-114, and R-397-000-227-112, the AME Young Investigator Research Grant R-279-000-546-305 (A*STAR Grant No. A1884c0017), and the Singapore-MIT Alliance for Research and Technology (SMART) Ignition Grant R-279-000-572-592.

References

1. T. J. Wallin, J. Pikul and R. F. Shepherd, Nat. Rev. Mater., 2018, 3, 84–100.
2. C. Wang, K. Sim, J. Chen, H. Kim, Z. Rao, Y. Li, W. Chen, J. Song, R. Verduzco and C. Yu, Adv. Mater., 2018, 30, 1706695.
3. T. Chen, O. R. Bilal, K. Shea and C. Daraio, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 5698–5702.
4. S. I. Rich, R. J. Wood and C. Majidi, Nat. Electron., 2018, 1, 102–112.
5. C. Walsh, Nat. Rev. Mater., 2018, 3, 78–80.
6. M. Schaffner, J. A. Faber, L. Pianegonda, P. A. Rühs, F. Coulter and A. R. Studart, Nat. Commun., 2018, 9, 878.
7. W. Xu and D. H. Gracias, ACS Nano, 2019, 13, 4883–4892.
8. N. W. Bartlett, M. T. Tolley, J. T. B. Overvelde, J. C. Weaver, B. Mosadegh, K. Bertoldi, G. M. Whitesides and R. J. Wood, Science, 2015, 349, 161–165.
9. L. Zhao, B. Bhatia, S. Yang, E. Strobach, L. A. Weinstein, T. A. Cooper, G. Chen and E. N. Wang, ACS Nano, 2019, 137(7), 7508–7516.
10. Y. Alapan, O. Yasa, O. Schauer, J. Giltinan, A. F. Tabak, V. Sourjik and M. Sitti, Sci. Robot., 2018, 3, eaar4423.
11. J. J. Huaroto, E. Suarez, H. I. Krebs, P. D. Marasco and E. A. Vela, IEEE Robot. Autom. Lett., 2019, 4, 17–24.
12. J. Yuan, W. Neri, C. Zakri, P. Merzeau, K. Kratz, A. Lendlein and P. Poulin, Science, 2019, 365, 155–158.
13. H. Zhao, K. O’Brien, S. Li and R. F. Shepherd, Sci. Robot., 2016, 1, eaai7529.
14. M. Rossi, M. Bianchi, E. Battaglia, M. G. Catalano and A. Bicchi, IEEE Trans. Biomed. Eng., 2019, 66, 138–149.
15. M. Cianchetti, C. Laschi, A. Menciassi and P. Dario, Nat. Rev. Mater., 2018, 3, 143–153.
16. W. Hu, G. Z. Lum, M. Mastrangeli and M. Sitti, Nature, 2018, 554, 81.
17. M. Sitti, Nat. Rev. Mater., 2018, 3, 74–75.
18. T.-H. Chang, Y. Tian, C. Li, X. Gu, K. Li, H. Yang, P. Sanghani, C. M. Lim, H. Ren and P.-Y. Chen, ACS Appl. Mater. Interfaces, 2019, 11, 10226–10236.
19. X. Le, W. Lu, J. Zhang and T. Chen, Adv. Sci., 2019, 6, 1801584.
20. K.-I. Jang, K. Li, H. U. Chung, S. Xu, H. N. Jung, Y. Yang, J. W. Kwak, H. H. Jung, J. Song, C. Yang, A. Wang, Z. Liu, J. Y. Lee, B. H. Kim, J.-H. Kim, J. Lee, Y. Yu, B. J. Kim, H. Jang, K. J. Yu, J. Kim, J. W. Lee, J.-W. Jeong, Y. M. Song, Y. Huang, Y. Zhang and J. A. Rogers, Nat. Commun., 2017, 8, 15894.
21. K. Li, X. Cheng, F. Zhu, L. Li, Z. Xie, H. Luan, Z. Wang, Z. Ji, H. Wang, F. Liu, Y. Xue, C. Jiang, X. Feng, L. Li, J. A. Rogers, Y. Huang and Y. Zhang, Adv. Funct. Mater., 2019, 29, 1806630.
22. Y. Zhang, A. D. Mickle, P. Gutruf, L. A. McIlvried, H. Guo, Y. Wu, J. P. Golden, Y. Xue, J. G. Grajales-Reyes, X. Wang, S. Krishnan, Y. Xie, D. Peng, C.-J. Su, F. Zhang, J. T. Reeder, S. K. Vogt, Y. Huang, J. A. Rogers and R. W. Gereau, Sci. Adv., 2019, 5, eaaw5296.
23. H. Wang, M. Totaro and L. Beccai, Adv. Sci., 2018, 5, 1800541.
24. A. F. Ruiz-Olaya, A. Lopez-Delis and A. F. da Rocha, in Handbook of Biomechatronics, ed. J. Segil, Academic Press, Cambridge, MA, 2019, ch. 8, pp. 283–317.
25. H. Yang, B. S. Yeow, T.-H. Chang, K. Li, F. Fu, H. Ren and P.-Y. Chen, ACS Nano, 2019, 13, 5410–5420.
26. Z. Zhai, Y. Wang and H. Jiang, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 2032–2037.
27. C. E. Machnicki, F. Fu, L. Jing, P.-Y. Chen and I. Y. Wong, J. Mater. Chem. B, 2019 10.1039/C9TB01006H.
28. V. Georgakilas, J. N. Tiwari, K. C. Kemp, J. A. Perman, A. B. Bourlinos, K. S. Kim and R. Zboril, Chem. Rev., 2016, 116, 5464–5519.
29. K. K. Kim, H. S. Lee and Y. H. Lee, Chem. Soc. Rev., 2018, 47, 6342–6369.
30. L. Jing, H. Li, R. Y. Tay, B. Sun, S. H. Tsang, O. Cometto, J. Lin, E. H. T. Teo and A. I. Y. Tok, ACS Nano, 2017, 11, 3742–3751.
31. L. Jing, M. K. Samani, B. Liu, H. Li, R. Y. Tay, S. H. Tsang, O. Cometto, A. Nylander, J. Liu, E. H. T. Teo and A. I. Y. Tok, ACS Appl. Mater. Interfaces, 2017, 9, 14555–14560.
32. K. Huang, Z. Li, J. Lin, G. Han and P. Huang, Chem. Soc. Rev., 2018, 47, 5109–5124.
33. J. Pang, R. G. Mendes, A. Bachmatiuk, L. Zhao, H. Q. Ta, T. Gemming, H. Liu, Z. Liu and M. H. Rummeli, Chem. Soc. Rev., 2019, 48, 72–133.
34. X. Zhang, Z. Lai, Q. Ma and H. Zhang, Chem. Soc. Rev., 2018, 47, 3301–3338.
35. F. Xia, H. Wang, J. C. M. Hwang, A. H. C. Neto and L. Yang, Nat. Rev. Phys., 2019, 1, 306–317.
36. Q. Zhang, J. Zhang, S. Wan, W. Wang and L. Fu, Adv. Funct. Mater., 2018, 28, 1802500.
37. M. Acerce, E. K. Akdoğan and M. Chhowalla, Nature, 2017, 549, 370.
38. T.-H. Chang, K. Li, H. Yang and P.-Y. Chen, Adv. Mater., 2018, 30, 1802418.
39. P.-Y. Chen, M. Zhang, M. Liu, I. Y. Wong and R. H. Hurt, ACS Nano, 2018, 12, 234–244.
40. P.-Y. Chen, M. Liu, Z. Wang, R. H. Hurt and I. Y. Wong, Adv. Mater., 2017, 29, 1605096.
41. P.-Y. Chen, J. Sodhi, Y. Qiu, T. M. Valentin, R. S. Steinberg, Z. Wang, R. H. Hurt and I. Y. Wong, Adv. Mater., 2016, 28, 3564–3571.
42. https://www.themanufacturer.com/articles/the-role-of-human-factors-in-the-future-of-manufacturing/, accessed August 2019.
43. https://grabcad.com/library/paxcyborg-backbone-1, accessed August 2019.
44. http://appshopper.com/mac/medical/arm-muscles-motion, accessed August 2019.
45. N. T. Jafferis, E. F. Helbling, M. Karpelson and R. J. Wood, Nature, 2019, 570, 491–495.
46. Y. Tu, J. Yuan, D. Lei, H. Tan, J. Wei, W. Huang and L. Zhang, J. Mater. Chem. A, 2018, 6, 8238–8243.
47. L. Zhang, P. e. Naumov, X. Du, Z. Hu and J. Wang, Adv. Mater., 2017, 29, 1702231.
48. X. Cai, Y. Luo, B. Liu and H.-M. Cheng, Chem. Soc. Rev., 2018, 47, 6224–6266.
49. H. Wu and M. R. Kessler, ACS Appl. Mater. Interfaces, 2015, 7, 5915–5926.
50. C. Yu, Q. Zhang, J. Zhang, R. Geng, W. Tian, X. Fan and Y. Yao, ACS Appl. Nano Mater., 2018, 1, 4875–4883.
51. W. Sheng, I. Amin, C. Neumann, R. Dong, T. Zhang, E. Wegener, W.-L. Chen, P. Förster, H. Q. Tran, M. Löffler, A. Winter, R. D. Rodriguez, E. Zschech, C. K. Ober, X. Feng, A. Turchanin and R. Jordan, Small, 2019, 15, 1805228.
52. M. Cui, S. Ren, H. Zhao, Q. Xue and L. Wang, Chem. Eng. J., 2018, 335, 255–266.
53. Z. Xia, G. Liu, Y. Dong and Y. Zhang, Prog. Org. Coat., 2019, 133, 154–160.
54. Y.-Z. Zhang, K. H. Lee, D. H. Anjum, R. Sougrat, Q. Jiang, H. Kim and H. N. Alshareef, Sci. Adv., 2018, 4, eaat0098.
55. B. Anasori, M. R. Lukatskaya and Y. Gogotsi, Nat. Rev. Mater., 2017, 2, 16098.
56. C. S. Boland, U. Khan, G. Ryan, S. Barwich, R. Charifou, A. Harvey, C. Backes, Z. Li, M. S. Ferreira, M. E. Möbius, R. J. Young and J. N. Coleman, Science, 2016, 354, 1257–1260.
57. F. Xiao, S. Naficy, G. Casillas, M. H. Khan, T. Katkus, L. Jiang, H. Liu, H. Li and Z. Huang, Adv. Mater., 2015, 27, 7196–7203.
58. J. Wang, Y. Wu, Y. Xue, D. Liu, X. Wang, X. Hu, Y. Bando and W. Lei, J. Mater. Chem. C, 2018, 6, 1363–1369.
59. D. Li, M. B. Müller, S. Gilje, R. B. Kaner and G. G. Wallace, Nat. Nanotechnol., 2008, 3, 101.
60. J. Chen, B. W. Yao, C. Li and G. Q. Shi, Carbon, 2013, 64, 225–229.
61. Y. Yang, Y. Tan, X. Wang, W. An, S. Xu, W. Liao and Y. Wang, ACS Appl. Mater. Interfaces, 2018, 10, 7688–7692.
62. C. Yang, Z. Liu, C. Chen, K. Shi, L. Zhang, X.-J. Ju, W. Wang, R. Xie and L.-Y. Chu, ACS Appl. Mater. Interfaces, 2017, 9, 15758–15767.
63. E. Wang, M. S. Desai and S.-W. Lee, Nano Lett., 2013, 13, 2826–2830.
64. W. Jiang, D. Niu, H. Liu, C. Wang, T. Zhao, L. Yin, Y. Shi, B. Chen, Y. Ding and B. Lu, Adv. Funct. Mater., 2014, 24, 7598–7604.
65. D. Niu, W. Jiang, H. Liu, T. Zhao, B. Lei, Y. Li, L. Yin, Y. Shi, B. Chen and B. Lu, Sci. Rep., 2016, 6, 27366.
66. L. Zong, X. Li, X. Han, L. Lv, M. Li, J. You, X. Wu and C. Li, ACS Appl. Mater. Interfaces, 2017, 9, 32280–32289.
67. J. Mu, C. Hou, G. Wang, X. Wang, Q. Zhang, Y. Li, H. Wang and M. Zhu, Adv. Mater., 2016, 28, 9491–9497.
68. T.-H. Chang, Y. Tian, D. L. Y. Wee, H. Ren and P.-Y. Chen, Small, 2018, 14, 1800596.
69. W. Wang, C. Xiang, Q. Zhu, W. Zhong, M. Li, K. Yan and D. Wang, ACS Appl. Mater. Interfaces, 2018, 10, 27215–27223.
70. K. Choi, S. Nam, Y. Lee, M. Lee, J. Jang, S. J. Kim, Y. J. Jeong, H. Kim, S. Bae, J.-B. Yoo, S. M. Cho, J.-B. Choi, H. K. Chung, J.-H. Ahn, C. E. Park and B. H. Hong, ACS Nano, 2015, 9, 5818–5824.
71. D. Pierleoni, Z. Y. Xia, M. Christian, S. Ligi, M. Minelli, V. Morandi, F. Doghieri and V. Palermo, Carbon, 2016, 96, 503–512.
72. J. Liu, R. G. Kutty, Q. Zheng, V. Eswariah, S. Sreejith and Z. Liu, Small, 2017, 13, 1602456.
73. J. S. Bunch, S. S. Verbridge, J. S. Alden, A. M. van der Zande, J. M. Parpia, H. G. Craighead and P. L. McEuen, Nano Lett., 2008, 8, 2458–2462.
74. F. Guo, G. Silverberg, S. Bowers, S.-P. Kim, D. Datta, V. Shenoy and R. H. Hurt, Environ. Sci. Technol., 2012, 46, 7717–7724.
75. R. R. Nair, H. A. Wu, P. N. Jayaram, I. V. Grigorieva and A. K. Geim, Science, 2012, 335, 442–444.
76. P. Y. Chen, M. Zhang, M. Liu, I. Y. Wong and R. H. Hurt, ACS Nano, 2018, 12, 234–244.
77. S. Lee, I. Jo, S. Kang, B. Jang, J. Moon, J. B. Park, S. Lee, S. Rho, Y. Kim and B. H. Hong, ACS Nano, 2017, 11, 5318–5324.
78. H. An, T. Habib, S. Shah, H. Gao, M. Radovic, M. J. Green and J. L. Lutkenhaus, Sci. Adv., 2018, 4, eaaq0118.
79. H. Guo, C. Lan, Z. Zhou, P. Sun, D. Wei and C. Li, Nanoscale, 2017, 9, 6246–6253.
80. H. Sun, Z. Xu and C. Gao, Adv. Mater., 2013, 25, 2554–2560.
81. W. Xu, S. K. Paidi, Z. Qin, Q. Huang, C.-H. Yu, J. V. Pagaduan, M. J. Buehler, I. Barman and D. H. Gracias, Nano Lett., 2019, 19, 1409–1417.
82. T. Deng, C. Yoon, Q. Jin, M. Li, Z. Liu and D. H. Gracias, Appl. Phys. Lett., 2015, 106, 203108.
83. W. Xu, Z. Qin, C.-T. Chen, H. R. Kwag, Q. Ma, A. Sarkar, M. J. Buehler and D. H. Gracias, Sci. Adv., 2017, 3, e1701084.
84. K. Pan, Y. Fan, T. Leng, J. Li, Z. Xin, J. Zhang, L. Hao, J. Gallop, K. S. Novoselov and Z. Hu, Nat. Commun., 2018, 9, 5197.
85. A. Sarycheva, A. Polemi, Y. Liu, K. Dandekar, B. Anasori and Y. Gogotsi, Sci. Adv., 2018, 4, eaau0920.
86. X. Zhang, J. Grajal, J. L. Vazquez-Roy, U. Radhakrishna, X. Wang, W. Chern, L. Zhou, Y. Lin, P.-C. Shen, X. Ji, X. Ling, A. Zubair, Y. Zhang, H. Wang, M. Dubey, J. Kong, M. Dresselhaus and T. Palacios, Nature, 2019, 566, 368–372.
87. H. Li, L. Jing, Z. L. Ngoh, R. Y. Tay, J. Lin, H. Wang, S. H. Tsang and E. H. T. Teo, ACS Appl. Mater. Interfaces, 2018, 10, 41707–41716.
88. X. Hou, Y. Chen, L. Lv, W. Dai, S. Zhao, Z. Wang, L. Fu, C.-T. Lin, N. Jiang and J. Yu, ACS Appl. Nano Mater., 2019, 2, 360–368.
89. J.-Y. Hong, S. Yun, J. J. Wie, X. Zhang, M. S. Dresselhaus, J. Kong and H. S. Park, Nanoscale, 2016, 8, 12900–12909.
90. Q. Zhang, X. Xu, H. Li, G. Xiong, H. Hu and T. S. Fisher, Carbon, 2015, 93, 659–670.
91. Z. Li, S. Wu, W. Lv, J.-J. Shao, F. Kang and Q.-H. Yang, Small, 2016, 12, 2674–2688.
92. R. Boston, A. Bell, V. P. Ting, A. T. Rhead, T. Nakayama, C. F. J. Faul and S. R. Hall, CrystEngComm, 2015, 17, 6094–6097.
93. Y. Saito, X. Luo, C. Zhao, W. Pan, C. Chen, J. Gong, H. Matsumoto, J. Yao and H. Wu, Adv. Funct. Mater., 2015, 25, 5683–5690.
94. H. Lee, C. Xia and N. X. Fang, Soft Matter, 2010, 6, 4342–4345.
95. D. Rus and M. T. Tolley, Nat. Rev. Mater., 2018, 3, 101–112.
96. R. Liang, Z. Li, L.-T. Weng, L. Zhang and G. Sun, Mater. Chem. Front., 2018, 2, 2076–2080.
97. W. Zhou, S. Qi, C. Tu and H. Zhao, J. Appl. Polym. Sci., 2007, 104, 2478–2483.
98. G. Alici, MRS Adv., 2018, 3, 1557–1568.
99. D. Avraham, E. Faran and D. Shilo, J. Phys. Commun., 2018, 2, 015015.
100. B. Jin, H. Song, R. Jiang, J. Song, Q. Zhao and T. Xie, Sci. Adv., 2018, 4, eaao3865.
101. T. Giffney, E. Bejanin, A. S. Kurian, J. Travas-Sejdic and K. Aw, Sens. Actuators, A, 2017, 259, 44–49.
102. T. Yamada, Y. Hayamizu, Y. Yamamoto, Y. Yomogida, A. Izadi-Najafabadi, D. N. Futaba and K. Hata, Nat. Nanotechnol., 2011, 6, 296–301.
103. C. Lee, X. Wei, J. W. Kysar and J. Hone, Science, 2008, 321, 385–388.
104. H. Hu, Z. Zhao, W. Wan, Y. Gogotsi and J. Qiu, Adv. Mater., 2013, 25, 2219–2223.
105. A. Ciarrocchi, A. Avsar, D. Ovchinnikov and A. Kis, Nat. Commun., 2018, 9, 919.
106. S. Cui, H. Pu, S. A. Wells, Z. Wen, S. Mao, J. Chang, M. C. Hersam and J. Chen, Nat. Commun., 2015, 6, 8632.
107. H. Yoo, R. Engelke, S. Carr, S. Fang, K. Zhang, P. Cazeaux, S. H. Sung, R. Hovden, A. W. Tsen, T. Taniguchi, K. Watanabe, G.-C. Yi, M. Kim, M. Luskin, E. B. Tadmor, E. Kaxiras and P. Kim, Nat. Mater., 2019, 18, 448–453.
108. T.-H. Chang, T. Zhang, H. Yang, K. Li, Y. Tian, J. Y. Lee and P.-Y. Chen, ACS Nano, 2018, 12, 8048–8059.
109. E. M. Alexeev, D. A. Ruiz-Tijerina, M. Danovich, M. J. Hamer, D. J. Terry, P. K. Nayak, S. Ahn, S. Pak, J. Lee, J. I. Sohn, M. R. Molas, M. Koperski, K. Watanabe, T. Taniguchi, K. S. Novoselov, R. V. Gorbachev, H. S. Shin, V. I. Fal’ko and A. I. Tartakovskii, Nature, 2019, 567, 81–86.
110. L. Qiu, Z. He and D. Li, Adv. Mater., 2018, 30, 1704850.
111. T. Cheng, G. Li, Y. Liang, M. Zhang, B. Liu, T.-W. Wong, J. Forman, M. Chen, G. Wang, Y. Tao and T. Li, Smart Mater. Struct., 2018, 28, 015019.
112. H. Li, L. Jing, W. Liu, J. Lin, R. Y. Tay, S. H. Tsang and E. H. T. Teo, ACS Nano, 2018, 12, 1262–1272.
113. S.-Y. Xie, Y. Wang and X.-B. Li, Adv. Mater., 2019, 31, 1900392.
114. X. Gao, H. Liu, D. Wang and J. Zhang, Chem. Soc. Rev., 2019, 48, 908–936.
115. A. Hirsch and F. Hauke, Angew. Chem., Int. Ed., 2018, 57, 4338–4354.
116. S. Presolski and M. Pumera, Mater. Today, 2016, 19, 140–145.
117. A. J. Mannix, B. Kiraly, M. C. Hersam and N. P. Guisinger, Nat. Rev. Chem., 2017, 1, 0014.
118. X. Wu, W. Liu, C. Zhang and L. Ren, Eur. Polym. J., 2018, 100, 12–17.
119. D. W. Kim, H. S. Jeong, K. O. Kwon, J. M. Ok, S. M. Kim and H.-T. Jung, Adv. Mater. Interfaces, 2016, 3, 1600534.
120. K.-G. Zhou, M. Zhao, M.-J. Chang, Q. Wang, X.-Z. Wu, Y. Song and H.-L. Zhang, Small, 2015, 11, 694–701.
121. C. Backes, R. J. Smith, N. McEvoy, N. C. Berner, D. McCloskey, H. C. Nerl, A. O’Neill, P. J. King, T. Higgins, D. Hanlon, N. Scheuschner, J. Maultzsch, L. Houben, G. S. Duesberg, J. F. Donegan, V. Nicolosi and J. N. Coleman, Nat. Commun., 2014, 5, 4576.
122. H. Song, J. Liu, B. Liu, J. Wu, H.-M. Cheng and F. Kang, Joule, 2018, 2, 442–463.
123. M. Shtein, R. Nadiv, M. Buzaglo, K. Kahil and O. Regev, Chem. Mater., 2015, 27, 2100–2106.
124. J. H. Kim, J. H. Jeong, N. Kim, R. Joshi and G.-H. Lee, J. Phys. D: Appl. Phys., 2018, 52, 083001.
125. M. Zhao, D.-B. Xiong, Z. Tan, G. Fan, Q. Guo, C. Guo, Z. Li and D. Zhang, Scr. Mater., 2017, 139, 44–48.
126. X. Li, Z. Li and Y.-W. Yang, Adv. Mater., 2018, 30, 1800177.
127. R. Liu, X. Kuang, J. Deng, Y.-C. Wang, A. C. Wang, W. Ding, Y.-C. Lai, J. Chen, P. Wang, Z. Lin, H. J. Qi, B. Sun and Z. L. Wang, Adv. Mater., 2018, 30, 1705195.
128. D. Barpuzary, K. Kim and M. J. Park, ACS Nano, 2019, 13, 3953–3963.
129. X. Feng and A. D. Schlüter, Angew. Chem., Int. Ed., 2018, 57, 13748–13763.
130. F. Yang, S. Cheng, X. Zhang, X. Ren, R. Li, H. Dong and W. Hu, Adv. Mater., 2018, 30, 1702415.
131. L. Jing, H. Li, R. Y. Tay, J. Lin, S. H. Tsang, E. H. T. Teo and A. I. Y. Tok, Macromol. Mater. Eng., 2018, 1800024.
132. R. J. Young, M. Liu, I. A. Kinloch, S. Li, X. Zhao, C. Vallés and D. G. Papageorgiou, Compos. Sci. Technol., 2018, 154, 110–116.
133. K. Chu, X.-h. Wang, F. Wang, Y.-b. Li, D.-j. Huang, H. Liu, W.-l. Ma, F.-x. Liu and H. Zhang, Carbon, 2018, 127, 102–112.
134. C. Yu, W. Gong, W. Tian, Q. Zhang, Y. Xu, Z. Lin, M. Hu, X. Fan and Y. Yao, Compos. Sci. Technol., 2018, 160, 199–207.
135. A. D. Dillon, M. J. Ghidiu, A. L. Krick, J. Griggs, S. J. May, Y. Gogotsi, M. W. Barsoum and A. T. Fafarman, Adv. Funct. Mater., 2016, 26, 4162–4168.
136. L.-H. Zeng, S.-H. Lin, Z.-J. Li, Z.-X. Zhang, T.-F. Zhang, C. Xie, C.-H. Mak, Y. Chai, S. P. Lau, L.-B. Luo and Y. H. Tsang, Adv. Funct. Mater., 2018, 28, 1705970.
137. S. Huang, G. He, C. Yang, J. Wu, C. Guo, T. Hang, B. Li, C. Yang, D. Liu, H.-J. Chen, Q. Wu, X. Gui, S. Deng, Y. Zhang, F. Liu and X. Xie, ACS Appl. Mater. Interfaces, 2019, 11, 1294–1302.
138. C. Kim, J.-C. Park, S. Y. Choi, Y. Kim, S.-Y. Seo, T.-E. Park, S.-H. Kwon, B. Cho and J.-H. Ahn, Small, 2018, 14, 1704116.
139. S.-Y. Cho, S. J. Kim, Y. Lee, J.-S. Kim, W.-B. Jung, H.-W. Yoo, J. Kim and H.-T. Jung, ACS Nano, 2015, 9, 9314–9321.
140. B. Han, Y.-L. Zhang, Q.-D. Chen and H.-B. Sun, Adv. Funct. Mater., 2018, 28, 1802235.
141. W. Ning, Z. Wang, P. Liu, D. Zhou, S. Yang, J. Wang, Q. Li, S. Fan and K. Jiang, Carbon, 2018, 139, 1136–1143.
142. B.-C. Min, R. Parasuraman, S. Lee, J.-W. Jung and E. T. Matson, Robot. Auton. Syst., 2018, 101, 57–73.
143. X. Pu, M. Liu, X. Chen, J. Sun, C. Du, Y. Zhang, J. Zhai, W. Hu and Z. L. Wang, Sci. Adv., 2017, 3, e1700015.

---