Magnetic assembly of transparent and conducting graphene-based functional composites

Abstract

Innovative methods producing transparent and flexible electrodes are highly sought in modern optoelectronic applications to replace metal oxides, but available solutions suffer from drawbacks such as brittleness, unaffordability and inadequate processability. Here we propose a general, simple strategy to produce hierarchical composites of functionalized graphene in polymeric matrices, exhibiting transparency and electron conductivity. These are obtained through protein-assisted functionalization of graphene with magnetic nanoparticles, followed by magnetic-directed assembly of the graphene within polymeric matrices undergoing sol-gel transitions. By applying rotating magnetic fields or magnetic moulds, both graphene orientation and distribution can be controlled within the composite. Importantly, by using magnetic virtual moulds of predefined meshes, graphene assembly is directed into double-percolating networks, reducing the percolation threshold and enabling combined optical transparency and electrical conductivity not accessible in single-network materials. The resulting composites open new possibilities on the quest of transparent electrodes for photovoltaics, organic light-emitting diodes and stretchable optoelectronic devices.

ARTICLE

Received 3 Dec 2015 | Accepted 27 May 2016 | Published 29 Jun 2016

Innovative methods producing transparent and exible electrodes are highly sought in modern optoelectronic applications to replace metal oxides, but available solutions suffer from drawbacks such as brittleness, unaffordability and inadequate processability. Here we propose a general, simple strategy to produce hierarchical composites of functionalized graphene in polymeric matrices, exhibiting transparency and electron conductivity. These are obtained through protein-assisted functionalization of graphene with magnetic nanoparticles, followed by magnetic-directed assembly of the graphene within polymeric matrices undergoing solgel transitions. By applying rotating magnetic elds or magnetic moulds, both graphene orientation and distribution can be controlled within the composite. Importantly, by using magnetic virtual moulds of predened meshes, graphene assembly is directed into double-percolating networks, reducing the percolation threshold and enabling combined optical transparency and electrical conductivity not accessible in single-network materials. The resulting composites open new possibilities on the quest of transparent electrodes for photovoltaics, organic light-emitting diodes and stretchable optoelectronic devices.

DOI: 10.1038/ncomms12078 OPEN

Magnetic assembly of transparent and conducting graphene-based functional composites

Hortense Le Ferrand1, Sreenath Bolisetty2, Ahmet F. Demirrs1, Rafael Libanori1, Andr R. Studart1

& Raffaele Mezzenga2

1 Complex Materials, Department of Materials, ETH Zurich, Zurich 8093, Switzerland. 2 Food and Soft Materials, Department of Health Sciences and Technologies, ETH Zurich, Zurich 8092, Switzerland. Correspondence and requests for materials should be addressed to A.R.S. (email: mailto:[email protected]

Web End [email protected] ) or to R.M. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12078

The increasing demand for optoelectronic tools from tissue-compatible biomedical devices for health monitoring to indiumtin-oxide-free electrodes for exible solar

cells has recently stimulated the search for cost-effective materials with comparable performance, but improved exibility to replace the commonly used stiff and brittle components14. There is an extensive library of existing potential materials combining electrical conductivity and transparency, mostly based on substrate lms coated with a mesh of nanocarbons or metal nanowires5,6. These materials present high electrical conductivity, corresponding to a sheet resistance down to 128 Ohm sq 1 with 95% transparency, while exibility is provided to a certain extent by the polymeric substrate. However, bulk materials containing mixtures of the conductive elements and the exible matrix present the advantage to provide more mechanical integrity over sandwich or layered structures, as well as the possibility to create more complex three-dimensional (3D) circuits. In addition, direct ink writing or 3D printing techniques can be applied to build on-demand materials with inks containing both the transparent support and the conductive elements79. Conductive and transparent polymers and blends based on poly(3,4-ethylenedioxythiophene) doped with polystyrene sulfonic acid (PEDOT:PSS) can reach transparency up to 87% combined with a conductivity of 1.35 S cm 1 for lms of 1 mm thickness10.

Yet, even in this case, good performances are typically recorded for ultrathin lms and based on the intrinsic conductivity of its constituents. Polymer composites, on the other hand, are attractive alternatives since the insulating polymeric matrix can be highly exible and optically transparent, whereas the electrical properties can be manipulated by controlling the architecture of conductive inclusions. The performance of the composite is then uniquely based not only on the intrinsic properties of its constituents but also on the arrangement of the conducting ller particles within the matrix and the contact between them. Although most physical properties of composite materials are generally enhanced by increasing the ller content, optical transparency tends to be reduced due to the difference in refractive index between ller particles and matrix. Thus, new approaches to fabricate exible composite materials displaying both enhanced electrical conductivity and high optical transparency must be developed to full the current requirements of optoelectronic applications.

Recent efforts have focused on employing conductive anisotropic particles exhibiting high aspect ratio to reduce the amount of opaque material needed to reach the percolation threshold in the composite1115. Elongated allotropes of carbon such as multi- or single-wall carbon nanotubes have been used to yield electrical conductivity corresponding to a sheet resistance down to 290 Ohm sq 1 in cellulose nanobrils aerogel membranes while maintaining 90% in transparency and 5% in strain at rupture16 or 34 Ohm sq 1 in 81% transparent polyethylene terephthalate-congo red single-wall carbon nanotube composites17. Graphene nanosheets are particularly promising in view of their unique mechanical and electronic properties1821 and their availability through well-established exfoliation methods2226. Nevertheless, even the remarkable percolation thresholds typically achieved in graphene-based composites are not low enough to yield materials with acceptable optical transparency and colour neutrality27. The use of the more processable graphene oxide may partially improve this scenario, yet reducing the conductivity, with the fate of the nal materials optoelectronic properties still depending on the initial precursors formulation and the processing route followed28. Although highly conducting and transparent materials based on graphene have already been obtained using processing pathways such as transfer

printing combined with thermal treatment29 or chemical vapour deposition followed by etching30,31, the fabrication costs associated with these techniques greatly reduce their potential for large-scale applications and call for new strategies.

One enticing approach to further decrease the amount of conductive particles needed to achieve the percolation threshold is to fabricate hierarchical composites exhibiting multiple percolating networks at different length scales32,33. While the potential of this approach has already been demonstrated for carbon nanotube-based composites34, translation of such concepts into industrially relevant processes requires simple and cost-effective strategies that allow for a deliberate control over the spatial distribution of the secondary network. Furthermore, combination of multiple percolation networks with sufcient optical transparency remains yet to be demonstrated.

In this study, we describe a simple processing route to fabricate transparent and conducting polymer-based composite lms using external magnetic elds to assemble magnetized graphene akes into hierarchical networks of predened features. Using virtual magnetic moulds, transparency is obtained by locally concentrating the dark conductive akes into continuous networks. When large micrometric mesh sizes are used as moulds, such control is achieved over areas spanning over several centimeters, resulting in double-percolating hierarchical networks. In addition to spatial control, tuning the orientation of the functionalized graphene akes within the microscopic pattern further enhances the electrical conductivity. By carefully engineering the mesh size of virtual magnetic moulds, we demonstrate how this new strategy can lead to transparent, electrically conducting polymergraphene composites with great potential for applications in advanced compliant optoelectronics.

ResultsSynthesis and processing strategy. Hydrophilic magnetically responsive reduced graphene oxide (m-rGO) akes are synthesized through decoration of exfoliated micrometric graphene oxide nanosheets (GO) with 10 nm superparamagnetic iron oxide nanoparticles (SPIONs; (Fig. 1ac)). The attachment of SPIONs is assisted by a multidomain protein, bovine serum albumin (BSA), which maintains the magnetic nanoparticles adsorbed on the ake surface throughout all the preparation steps. The physical adsorption of the BSA generates a predened molecular coating that settles the saturation limit to the number of adsorbed molecules and provides more available sites for the adsorption of SPIONs. After partial reduction by BSA, a treatment with hydrazine at 80 C converts the GO into reduced graphene oxide (rGO)35 (see Supplementary Fig. 1 and Supplementary Discussion for details about reduction). If suspended in a liquid, the prepared m-rGO akes exhibit high magnetic response, similar to the ultrahigh magnetic behaviour observed for other anisotropic particles decorated with SPIONs36. Remarkably, in contrast with the hydrophobic nature of rGO, the presence of polar organic groups in the surface-adsorbed SPIONs and BSA enables the successful dispersion of the m-rGO akes in polar solvents, preventing the formation of possible defects associated with graphitic aggregation, clustering or deterioration of the matrix during post-reduction treatments37,38. We therefore take an advantage of this unusual hydrophilicity of the m-rGO akes to incorporate them into two exemplary commonly used hydrogels, namely gelatin from bovine skin and poly(2-acrylamido-2-methyl-1-propanesulfonic acid) (PAMPS). These hydrogels exhibit a liquid-to-solid transition, which can be effectively used to x the orientation and spatial distribution of the m-rGO akes after the magnetically driven assembly process (see Fig. 1d,

2 NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12078 ARTICLE

a

Graphite Exfoliated graphene oxide

Exfoliated magnetic reduced graphene oxide (m-rGO)

Fe3O4

BSA

Hummers method

Hydrazine 80C, 20 h

b c

d

Control over the orientation

Control over the spatial distribution

Horizontal

H

Static magnet

Localization on the template

z

x

y

Vertical

55 C

55 C

Co or Ni template

Drying

Control of the orientation

Dispersion of m-rGO

Control of spatial distribution

Drying

Viscosity

Solid

S

Solid

Gelling

Gelling

Liquid Gelling

Solid

Gelling Solidd

Viscosity

25 55 25

25

55

25

55

Temperature (C)

Figure 1 | Preparation of graphene-based composite lms with predened architectures. (a) Synthesis of hydrophilic magnetically responsive reduced graphene oxide (m-rGO) mediated by bovine serum albumin (BSA). Transmission electron microscopic (TEM) images of (b) exfoliated graphene oxide and (c) m-rGO akes ((b,c) Scale bar, 1 mm) with 10-nm diameter iron oxide nanoparticles (Scale bar, 200 nm (inset)). (d) Processing of gelatin-based composites with magnetic control over the orientation or spatial distribution of the m-rGO akes. The magnetic assembly is performed in the liquid phase and is followed by consolidation of the matrix to yield composite materials with tailored structures at both nano- and microscales.

Supplementary Figs 2 and 3 and Supplementary Methods for details of the process and applicability of the strategy to polyurethanes).

The obtained m-rGO akes can be deliberately aligned using low-cost rare-earth magnets. Biaxial horizontal alignment of the m-rGO akes in the xy-plane is achieved by applying a horizontally rotating magnetic eld36. Literature reports that vertical orientation of anisotropic particles is nontrivial and is usually achieved by laborious and high power energy means or other costly post processing techniques3942. Here we deliberately

orient the m-rGO akes vertically by applying low static magnetic elds of 50 mT parallel to the z axis (Fig. 1d, left).

Besides orientation, the spatial distribution of m-rGO akes can be controlled using magnetic elds exhibiting gradients in magnitude. Indeed, local gradients in magnetic eld generate magnetophoretic forces that can efciently attract suspended particles exhibiting magnetic susceptibility different from that of the surrounding liquid43. Graded magnetic elds can be easily designed using small permanent magnets, magnetically patterned stripes or virtual magnetic moulds comprising a metallic

NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12078

template, typically nickel or cobalt, placed above a larger permanent magnet (Fig. 1d right)43,44. All these methods enable scalable localization of m-rGO akes in predetermined congurations and in time scales that cannot be achieved with 3D printing or other lithographic methods9,45.

Orientation and spatial distribution control. We demonstrate the ability to control the orientation and spatial distribution of m-rGO akes by Wide- and Small-Angle X-rays Scattering measurements and optical microscopy (Fig. 2; Supplementary Fig. 4). The signicant increase in the scattering intensity from composite lms containing m-rGO akes aligned parallel to the X-ray beam (along the z axis) compared with a horizontal conguration is the rst evidence of the successful alignment of the akes in the intended direction (Fig. 2a). In addition, in the vertical alignment, several intense Bragg reections emerge, characteristic of either gelatin, rGO or SPIONs (see Supplementary Discussion), which cannot be resolved in the horizontal alignment.

In the two examples in Fig. 2b,c we used magnetic elds to drive the spatial distribution of graphene akes within the xy-plane of the composite. In Fig. 2b, the m-rGO akes are assembled within

the gelatin matrix into a striated pattern using the low coercivity magnetic stripe of a standard train ticket. After consolidating the matrix, the resulting 30.7-mm-thick lm exhibits features of similar dimensions as the magnetic stripe, with a periodicity of 175 mm between m-rGO-free and m-rGO-concentrated areas. To showcase the potential of using virtual magnetic moulds to obtain m-rGO-based composites exhibiting more complex microstructural designs, we fabricated a template using a 250-mm-thick nickel wire bended and curved into a ower-like shape. Placing the metallic template 1 mm above a permanent magnet of 250 mT, a virtual magnetic mould is generated, featuring magnetic eld microgradients to guide the assembly of the m-rGO akes in the shape of the template. The mirrored picture of the template and the optical microscopy image of the resulting 0.02 vol% m-rGO-PAMPS composite lm conrm that shape and dimension of the virtual magnetic mould template are precisely translated to the m-rGO-based composite (Fig. 2c).

Tuning transparency. Positioning m-rGO akes into two-dimensional (2D) patterns through magnetic manipulation of diluted dispersions enables the creation of hierarchical double-percolating networks (Fig. 3). At the nanoscale the

a

p-I p-II p-III p-IV p-V

Incident beam

z

Vertical alignment

(pixel)

200

100

0y

200 600 1,000 1,400

I (q)

40

30

20

10

00.5 1.5 2.5 3.5

1 2 3 q (A1)

Horizontal alignment

x (pixel)

y

x

(pixel)

200

100

0y

200 600 1,000 1,400

x (pixel)

0 1 2 3 4 5

b

c

Graphene composite

200

100

Template

175 m

Grey scale

0 0.6 1.2 1.8 Distance (mm)

Figure 2 | Hierarchically structured composites exhibiting controlled orientation and spatial distribution of m-rGO akes. (a) Wide-angle X-ray Scattering (WAXS) of 0.96 vol% m-rGOgelatin composites exhibiting horizontal (black) and vertical (grey) orientation of graphene akes. Scattering peaks correspond to the gelatin matrix (p-I in brown), the SPIONs (p-II and p-V, blue) and the graphene (p-III and p-IV, red). (b,c) Spatial control of m-rGO akes (b) within a gelatin matrix (7.84 vol% m-rGO) using a patterned magnetic stripe (Scale bar, 2 mm) and (c) within a PAMPS matrix (0.037 vol% m-rGO) using a virtual magnetic mould with a complex shape (Scale bar, 2 mm).

4 NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12078 ARTICLE

a b

Width 7 m

Width 7 m

Width 50 m

Width 25 m

100

50%

30%

28%

9.5%

80

Transparency (%)

60

Optically transparent

Prediction

40

Template area (%)

20

0

20 40 80

60 100 Template area (%)

0

c

Without composite

With graphene composite

0 20 40 80

60 100

Template area (%)

Figure 3 | Optical transparency of hierarchical graphene-based composites assembled on mesh-like magnetic virtual moulds. (a) Optical micrographs of PAMPS composites containing 0.065 vol% m-rGO akes assembled into mesh-like patterns of decreasing template area and different line width (Scale bar, 100 mm). (b) The transparency of the 30-mm-thick composite lm containing 0.065 vol% m-rGO-PAMPS increases with decreasing coverage area of the templates. (c) Demonstration of the improved optical transparency by controlling the spatial distribution of m-rGO akes within 0.75 vol% rGOgelatin composite lms using metallic templates of different coverage areas (Scale bar, 1 mm).

network consists of an assembly of interconnected m-rGO akes, whereas the 2D pattern gives rise to the second network at the microscale. Differently from the great majority of the approaches proposed in literature, such a directed hierarchical architecture is an efcient strategy to control the optical transparency of the composite lm without altering its initial composition (see Supplementary Methods for more details about the denition of transparency adopted here). Similarly, colour neutrality can be preserved up to relatively high concentrations of graphene. To illustrate this approach, we designed a series of experiments where 0.065 vol% m-rGO akes in PAMPS are localized into continuous mesh patterns of decreasing coverage area using 10-mm-thick nickel transmission electron microscopic grids, directly affecting the overall transparency of the composite lm (Fig. 3a; Supplementary Fig. 5). The optical transparency of each composite shows a good agreement with the values expected for lms with akes covering precisely the templated area, conrming that the absolute template area controls the transparency of the composite, regardless of the line thickness of the template (Fig. 3b). The striking increase in optical transparency enabled by our approach is further evidenced by comparing composite lms containing homogeneously distributed akes with those fabricated using magnetic virtual moulds. For the same lm thickness of 472 mm, 0.75 vol% rGOgelatin composites with randomly distributed m-rGO akes are dark and opaque, whereas lms obtained by using a 28% area coverage template are 80% more transparent, allowing the observation of objects throughout the composite lm (Fig. 3c).

Combining transparency with electrical conductivity. Locally concentrating the conductive rGO nanosheets into continuous mesh patterns not only enhances the overall optical transparency

of the lm but also creates electrically conductive paths that increase the global conductivity of the composite material (Fig. 4). To study this unusual combination of properties in details, we rst investigate the local electrical conductivity of individual stripes made by magnetically driven localization of the m-rGO akes (one hierarchical level, Fig. 4a). Gelatin-based composites with different global concentrations of graphene akes concentrated in individual stripes were fabricated using a 100-mm-thick and 1-mm-wide cobalt ribbon as a virtual magnetic mould positioned 1 mm above a 250 mT permanent magnet. The global volume fraction of rGO, fi is calculated from the density of the m-rGO akes and the coverage of iron oxide nanoparticles and BSA (details in Supplementary Fig. 6 and Supplementary Methods). The local electrical conductivity is measured by contacting two electrodes on the stripe or in the matrix area (see Supplementary Fig. 7 and Supplementary Methods). For global concentrations of 0.45 vol% or lower, the local volume fraction of graphene is not sufciently high to form a percolating network and the composite remains insulating. Increasing the global concentration of m-rGO akes to 0.75 vol% results in composites with conducting lines and insulating surrounding matrix. Thus, the minimum volume fraction needed to obtain a local percolating network lies between 0.45 and 0.75 vol%. To the best of our knowledge, this is also the rst example of magnetized rGO composite exhibiting electrically conducting properties4648.

The combination of individual conductive paths at the nanoscale into percolating patterns at a coarser length scale is ideal for the electrical conductivity of the resulting hierarchical composite lms. Similar to 3D percolating networks formed in polymer melts33, localizing the conductive elements into continuous mesh-like patterns of different sizes leads to geometrically dened 2D hierarchical networks with lowered global percolating thresholds (two hierarchical levels, Fig. 4b).

NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12078

a

b

10 %

28 %

Stripe

Matrix

100 %

0 1 2

0.5 1.5

Template area:

c(10%)

c(28%)

c(100%)

[p10]i (%)

Global conductivity [afii9846](S.cm1 )

102

108

Local conductivity [afii9846](S.cm1 )

107

102

108

103

105

103

105

104

104

Stripe

Insulating

1

Insulating

[p10] c(vol%)

0.5

40 %

106

106

107

Matrix

0 0 50 100

Template area (%)

0 2

0.5

1 1.5

0 2

0.5

1 1.5

Global volume fraction of rGO [p10]i (%)

Global volume fraction of rGO [p10]i (%)

c d

Template area 100%

Compressive strain (%)

Template area 10%

Compressive strain (%)

2.97 vol%

[p10]i (%) =

3.922 vol%

[p10]i (%) = 0.89 vol%

3103

1103

5.348 vol%

Global conductivity[afii9846](S.cm1 )

Global conductivity[afii9846](S.cm1 )

8105

2103

6105

4105

2105

0

0 0 0.4 0.8 1.2

0

0.4 0.8 1.2 1.6

e Homogeneous rGO distribution Programmed rGO spatial distribution

1 1.5

0.5 2

Optically transparent

Electrically conductive

+

Template 10%

100

0

102

108

Global conductivity [afii9846](S.cm 1)

80

104

103

105

Optically transparent

Electrically conducting

Optically transparent

Transparency (%)

60

Electrically conducting

40

106

107

0 2 Global volume fraction of rGO (%)

20

0

0.5

1 1.5

Global volume fraction of rGO (%)

Figure 4 | Electric response and transparency of gelatin composite lms with magnetically driven percolation threshold. (a) Optical micrographs (Scale bar, 500 mm) and electrical conductivity measurements demonstrating the local control over electrical conductivity by assembling a volume fraction fi (vol%)

of m-rGO akes into a stripe in gelatin composite lms (one hierarchical level). (b) Reduction of the percolation threshold fc as a function of the moulds template area in gelatin composite lms (two hierarchical levels). Experimental global conductivity points are tted with equation (1). (c,d) Evolution of the global conductivity under compressive strain for composite lms prepared with a magnetic template area of 100% and 10%, respectively. The dotted lines are guides to the eyes underlining the trend. The inset in d shows a schematic drawing of the set-up, indicating the direction of the pattern and measuring electrodes relative to the applied force. (e) Controlling the rGO spatial distribution using a magnetic template of 10% area leads to transparent and electrically conductive gelatin lms (dark blue region, right) of otherwise opaque and insulating homogeneous lms for rGO global volume fractions of 0.650.85 vol% (region framed in grey, left). The optical micrographs were obtained from gelatin lms containing 0.75 vol% rGO (Scale bar, 500 mm).

6 NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12078 ARTICLE

Global electrical conductivities of the hierarchical networks are measured by depositing a goldpalladium electrode spanning the longest dimension of the templates on the surface of the composite lms directly facing the magnet. To rationalize the effect of the spatial distribution of m-rGO akes on the electrical properties of the lms, we applied the following percolation model49:

s C fi fct; for fi4fc; 1 where s is the global electrical conductivity, fi the global volume fraction of rGO in the composite lm, fc the percolating threshold, t the conductivity exponent describing the percolation behaviour above threshold and C an empirical constant. This equation is valid for the overall single percolation as well as for multiple percolated systems, as demonstrated both theoretically and experimentally32,33. Fitting equation (1) to the experimental conductivity values conrms the expected percolating behaviour of the system above the threshold. Exponents t of 1.76 and 1.95 were obtained for composite lms fabricated from templates with coverage areas of 10% and 28%, respectively. Such values are lower than the expected exponent of t2D network 2 t1D

network

2.7, where t1D network is the exponent in the case of a single network, equal to 1.35 in this system32,33 (Supplementary Table 1). This probably results from the combined detrimental inuence of the vertical orientation of the rGO nanosheets in composite lms with higher volume fractions of particles with the non-conductive BSA/SPIONs coating of the nanosheets, reducing the number of contact points (see Supplementary Note 1 for details). Well-established approaches to remove surface-adsorbed species after assembly could be used to improve further the conductivity while maintaining the hierarchical magnetic assembly. Despite the lower exponent t, a remarkable decrease of 40% of the percolation threshold is observed for composite lms containing m-rGO akes spatially distributed over a template area of 10% (see Supplementary Fig. 8, Supplementary Table 2 and Supplementary Discussion for details on conductivity measurements on the opposite face of the lms). This clearly demonstrates the potential of this method to tailor the percolation threshold of conducting anisotropic particles and to enhance the electrical conductivity of ller-loaded polymer lms.

A possible further advantage of the proposed strategy is that of allowing the production of conductive and stretchable functional composites serving as strain sensors of high precision. Indeed, thanks to the intrinsic exibility of polymers, we conducted in-plane conned compression experiments on polydimethylsiloxane substrates covered by the gelatin m-rGO lm while recording the change in electrical conductivity (details in Supplementary Methods and Fig. 4c,d). As the compression strain increases, the graphene akes are brought into closer contact, facilitating the electron transport and thus increasing the local conductivity. Remarkably, the benets of the magnetically tuned percolation threshold become apparent by comparing the onset of the strain detection: a comparable detection limit of 1.1% strain is reached using just 0.89 vol% m-rGO akes in lms exhibiting 10% template area (Fig. 4d), as opposed to the2.97 vol% needed in the case of a template area of 100% (Fig. 4c). In addition, above this detection limit, an increase of strain as small as 0.005% can be detected in 10% template area, in contrast with a minimum detectable strain of 0.16% for homogeneously distributed lms (for fi 5.35 vol%). Such a small resolution has

not yet been reported in stretchable and transparent strain sensors used for health-monitoring devices where high strains, typically up to 300%, is the main property targeted5052. In these systems, an error in strain of 2% is usually limiting the sensitivity53. Although other stretchable composites have been developed that reaches a detection limit down to 0.1% strain in

carbon black lled rubber54 or 0.5% in graphene and silver nanoparticles sandwich structures55, they do not possess the optical transparency required for optoelectronic devices. The high sensitivity provided by our patterned composites indicates that the hierarchical network of m-rGO akes not only decreases the global percolation threshold but also makes the conduction paths more strongly inuenced by applied deformations. Furthermore, the formation of a hierarchical network of modied graphene akes is also benecial to keep the high deformability of the host matrix, which otherwise becomes brittle when the akes are homogeneously distributed. These results underline the benets that our strategy provides in adapting the mesh pattern geometry and the matrix composition for advances in the eld of strain sensors, where high sensitivity is required in combination with the specic requirements and design found in health-monitoring systems and integrated electronic circuits.

Our ability to tune and control the spatial distribution of conductive paths as a means to combine optical transparency and electrical conductivity in the same composite is best illustrated when the obtained data are displayed simultaneously in a single plot of the relevant properties as a function of the global volume fraction of rGO (Fig. 4e). In composites containing homogenously distributed graphene akes the global volume fractions required to achieve optical transparency and electrical conductivity are mutually excluded. Instead, using an initial template area of 10%, rGOgelatin composite lms containing0.65 to 0.85 vol% of rGO can be made both transparent and conductive (Fig. 4e).

DiscussionWe have demonstrated that geometrically controlling the spatial distribution and orientation of magnetically functionalized graphene in polymer matrices is a simple and potentially up-scalable route for the design and fabrication of cost-effective, transparent and electrically conducting graphene-based functional composites. The design possibilities offered by the magnetic manipulation of graphene are mirrored by the great exibility on the format of virtual magnetic moulds, allowing the creation of complex shapes with high efciency and minimal efforts. Our results are complementary to existing approaches to design conductive and transparent materials, and prove the principle that conductive lms can made transparent through magnetic assembly even if highly light absorbing constituents such graphene are used as the conductive phase. The implementation of such approach to the wide range of polymer lms coated with metal nanowires and nanocarbon materials would enable a further increase in the electrical conductivity through the use of higher ller concentrations without compromising the lms optical transparency. Alternatively, the formation of a double percolation network of wires or nanotubes would enable the reduction of the ller content without sacricing the electrical conductivity of the lm.

This provides an exciting opportunity for the fabrication of next-generation stretchable optoelectronic sensors and devices that combine optimized design and materials meant to closely match target functions. Given the exibility of the substrate, the known stretchability of graphenepolymer composites and the biocompatibility of graphene56,57, potential applications that could benet from such systems range broadly from bio-integrated high-sensitivity strain sensors to high efciency conformable solar cells.

Methods

Preparation of magnetic-reduced graphene oxide. Graphene oxide was prepared following a documented protocol58 (see details in the Supplementary Methods).

NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12078

Spherical magnetite nanoparticles were synthesized by co-precipitation of FeSO4 and FeCl3 in the presence of NaOH. In a typical procedure, 0.25 mmol

FeSO4 7H2O and 0.5 mmol FeCl3 6H2O (both from Sigma-Aldrich) were

dissolved in 25 ml water and vigorously stirred. After heating to 60 C, 10 ml NaOH solution at 2 wt% was added with 250 mg of hexadecylpyridinium bromide (Fluka).

A measure of 18 ml of the GO stock solution was mixed with 2 ml of 0.5 wt% of the BSA protein (Sigma-Aldrich). In a following step, this suspension was stirred with 250 ml of 4.4 wt% SPIONs suspension for 1 h to allow for the physical adsorption of the SPIONs at the surface of the graphene oxide sheets through interactions with the BSA. Volume of 100 ml of reducing agent hydrazine monohydrate (Sigma-Aldrich) was added at 80 C for 20 h under continuous stirring. After cooling to room temperature, the mixtures were kept at 4 C.

Fabrication of m-rGOgelatin composites. A preheated (55 C, 30 min) aqueous solution of 20 wt% gelatin (from bovine skin, Sigma-Aldrich) was mixed with the corresponding volume of the suspension of m-rGO in water and stirred at 55 C. For the experiments with controlled orientation, the mixture was cast into small polyethylene moulds. A static vertical magnetic eld of 50 mT was used for the uniaxial alignment of the graphene akes along the z axis, whereas the samples with biaxially aligned akes were obtained by rotating the magnetic eld on the xy plane. The samples were afterwards cooled down to room temperature and dried for 24 h in ambient conditions to consolidate the gelatin and x the designed architecture (Supplementary Fig. 2). Samples with controlled spatial distribution of m-rGO akes were fabricated by casting onto a commercial magnetic stripe (train ticket) followed by cooling down and drying at ambient conditions (Fig. 2c). To demonstrate the decrease of the percolation threshold with the localization of the rGO, 250 ml of m-rGOgelatin mixtures were cast into a well with bare surface, or with a nickel or cobalt template covered by a Teon foil at the bottom (Supplementary Fig. 5). The wells were positioned on a permanent neodymium magnet of 250 mT (Supermagnete, Switzerland), before casting. The samples were then cooled to room temperature and dried in air, overnight.

Fabrication of m-rGO-PAMPS composites. PAMPS hydrogel was synthesized according to a modied process59 (see Supplementary Methods for more details). The monomer solution was mixed with the relevant amount of solution of m-rGO and deposited in similar wells as described previously. The nickel template conformed in the shape of a ower was made by twisting a nickel wire. The trio magnet, well and suspension were placed in a dark chamber and irradiated by ultraviolet light (OmniCure Series 1000, Lumen Dynamics) for 2 min at 60% of the maximum power. The gelled composite lms were then dried overnight at ambient temperature.

Characterization methods. Characterization of the m-rGO and the composites lm is detailed in the Supplementary Methods.

Data availability. The data that support the ndings of this study are available from the corresponding authors on request.

References

1. Rogers, J. A., Someya, T. & Huang, Y. Materials and mechanics for stretchable electronics. Science 327, 16031607 (2010).

2. Xu, S. et al. Soft microuidic assemblies of sensors, circuits, and radios for the skin. Science 344, 7074 (2014).

3. Sam, F. L. M., Mills, C. A., Rozanski, L. J. & Silva, S. R. P. Thin lm hexagonal gold grids as transparent conducting electrodes in organic light emitting diodes. Laser Photon. Rev. 179, 172179 (2014).

4. Libanori, R. et al. Stretchable heterogeneous composites with extreme mechanical gradients. Nat. Commun. 3, 12651274 (2012).

5. Mallikarjuna, K., Hwang, H.-J., Chung, W.-H. & Kim, H.-S. Photonic welding of ultra-long copper nanowire network for exible transparent electrodes using white ash light sintering. RSC Adv. 6, 47704779 (2016).

6. Kim, K. et al. Stretchable and transparent electrodes based on in-plane structures. Nanoscale 7, 1457714594 (2015).

7. Koga, H. et al. Transparent, conductive, and printable composites consisting of TEMPO-oxidized nanocellulose and carbon nanotube. Biomacromolecules 14, 11601165 (2013).

8. Secor, E. B., Prabhumirashi, P. L., Puntambekar, K., Geier, M. L. & Hersam, M.C. Inkjet printing of high conductivity, exible graphene patterns. J. Phys. Chem. Lett. 4, 13471351 (2013).9. Garca-Tunon, E. et al. Printing in Three dimensions with graphene. Adv. Mater. 27, 16881693 (2015).

10. Sun, J., Gerberich, W. W. & Francis, L. F. Transparent, conductive polymer blend coatings from latex-based dispersions. Prog. Org. Coatings 59, 115121 (2007).

11. Celzard, A. & March, J. F. Non-universal conductivity critical exponents in anisotropic percolating media: A new interpretation. Phys. A Stat. Mech. Appl. 317, 305312 (2003).

12. Li, J. & Kim, J. K. Percolation threshold of conducting polymer composites containing 3D randomly distributed graphite nanoplatelets. Compos. Sci. Technol. 67, 21142120 (2007).

13. Kyrylyuk, A. V. & van der Schoot, P. Continuum percolation of carbon nanotubes in polymeric and colloidal media. Proc. Natl Acad. Sci. USA 105, 82218226 (2008).

14. Moriarty, G. P., Whittemore, J. H., Sun, K. A., Rawlins, J. W. & Grunlan, J. C. Inuence of polymer particle size on the percolation threshold of electrically conductive latex-based composites. J. Polym. Sci. Part B Polym. Phys. 49, 15471554 (2011).

15. Li, D. & Kaner, R. B. Graphene-based materials. Science 320, 11701171 (2008).16. Toivonen, M. S., Kaskela, A., Rojas, O. J., Kauppinen, E. I. & Ikkala, O. Ambient-dried cellulose nanobril aerogel membranes with high tensile strength and their use for aerosol collection and templates for transparent, exible devices. Adv. Funct. Mater. 25, 66186626 (2015).

17. Sundramoorthy, A. K., Wang, Y.-C. & Gunasekaran, S. Low-temperature solution process for preparing exible transparent carbon nanotube lm for use in exible supercapacitors. Nano Res. 8, 34303445 (2015).

18. Lee, C., Wei, X., Kysar, J. W. & Hone, J. Measurement of the elastic properties and intrinsic strength of monolayer graphene. Science 321, 385388 (2008).

19. Nair, R. R. et al. Fine structure constant denes visual transparency of graphene. Science 320, 13081312 (2008).

20. Mayorov, A. S. et al. Micrometer-scale ballistic transport in encapsulated graphene at room temperature. Nano Lett. 11, 23962399 (2011).

21. Balandin, A. A. Thermal properties of graphene and nanostructured carbon materials. Nat. Mater. 10, 569581 (2011).

22. Stankovich, S. et al. Synthesis of graphene-based nanosheets via chemical reduction of exfoliated graphite oxide. Carbon 45, 15581565 (2007).

23. Park, S. & Ruoff, R. S. Chemical methods for the production of graphenes. Nat. Nanotechnol. 4, 217224 (2009).

24. Paton, K. R. et al. Scalable production of large quantities of defect-free few-layer graphene by shear exfoliation in liquids. Nat. Mater. 13, 624630 (2014).

25. Yi, M. & Shen, Z. A review on mechanical exfoliation for scalable production of graphene. J. Mater. Chem. A 3, 1170011715 (2015).

26. Schniepp, H. C. et al. Functionalized single graphene sheets derived from splitting graphite oxide. J. Phys. Chem. B 110, 85358539 (2006).

27. Hu, H. et al. Preparation and electrical conductivity of graphene/ultrahigh molecular weight polyethylene composites with a segregated structure. Carbon 50, 45964599 (2012).

28. Zheng, Q., Li, Z., Yang, J. & Kim, J.-K. Graphene oxide-based transparent conductive lms. Prog. Mater. Sci. 64, 200247 (2014).

29. Wang, S. J., Geng, Y., Zheng, Q. & Kim, J.-K. Fabrication of highly conducting and transparent graphene lms. Carbon 48, 18151823 (2010).

30. Kim, K. S. et al. Large-scale pattern growth of graphene lms for stretchable transparent electrodes. Nature 457, 706710 (2009).

31. Bae, S. et al. Roll-to-roll production of 30-inch graphene lms for transparent electrodes. Nat. Nanotechnol. 5, 574578 (2010).

32. Levon, K., Margolina, A. & Patashinsky, A. Z. Multiple percolation in conducting polymer blends. Macromolecules 26, 40614063 (1993).

33. Mezzenga, R. et al. Templating organic semi- conductors via self-assembly of polymer colloids. Science 299, 18721874 (2003).

34. Zhang, S., Deng, H., Zhang, Q. & Fu, Q. Formation of conductive networks with both segregated and double-percolated characteristic in conductive polymer composites with balanced properties. ACS Appl. Mater. Interfaces 6, 68356844 (2014).

35. Liu, J., Fu, S., Yuan, B., Li, Y. & Deng, Z. Toward a universal adhesive nanosheet for the assembly of multiple nanoparticles based on a protein-induced reduction/decoration of graphene oxide. J. Am. Chem. Soc. 132, 72797281 (2010).

36. Erb, R. M., Libanori, R., Rothfuchs, N. & Studart, A. R. Composites reinforced in three dimensions by using low magnetic elds. Science 335, 199204 (2012).

37. Noh, Y. J. et al. Ultra-high dispersion of graphene in polymer composite via solvent free fabrication and functionalization. Sci. Rep. 5, 91419148 (2015).

38. Cassagnau, P. Linear viscoelasticity and dynamics of suspensions andmolten polymers lled with nanoparticles of different aspect ratios. Polymer 54, 47624775 (2013).

39. Yoon, Y. et al. Vertical alignments of graphene sheets spatially and densely piled for fast ion diffusion in compact supercapacitors. ACS Nano 8, 45804590 (2014).

40. Wu, C., Li, F., Zhang, Y. & Guo, T. Field emission from vertical graphene sheets formed by screen-printing technique. Vacuum 94, 4852 (2013).

41. Yu, K. et al. Patterning vertically oriented graphene sheets for nanodevice applications. J. Phys. Chem. Lett. 2, 537542 (2011).

42. Cai, M. et al. Fast response, vertically oriented graphene nanosheet electric double layer capacitors synthesized from C2H2. ACS Nano 8, 58735882 (2014).

43. Demirrs, A. F., Pillai, P. P., Kowalczyk, B. & Grzybowski, B. A. Colloidal assembly directed by virtual magnetic moulds. Nature 503, 99103 (2013).

8 NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12078 ARTICLE

44. Libanori, R., Reusch, F. B., Erb, R. M. & Studart, A. R. Ultrahigh magnetically responsive microplatelets with tunable uorescence emission. Langmuir 29, 1467414680 (2013).

45. Zhang, Q. et al. Solution-processable graphene mesh transparent electrodes for organic solar cells. Nano Res. 6, 478484 (2013).

46. Yang, J. H., Ramaraj, B. & Yoon, K. R. Preparation and characterization of superparamagnetic graphene oxide nanohybrids anchored with Fe3O4

nanoparticles. J. Alloys Compd. 583, 128133 (2014).47. Myekhlai, M. et al. Synthesis and characterization of the graphene- Fe3O4-

hybrid composite. J. Nanosci. Nanotechnol. 15, 20472051 (2015).48. Yan, H., Wang, R., Li, Y. & Long, W. Thermal conductivity of magnetically aligned graphenepolymer composites with Fe3O4-decorated graphene nanosheets. J. Electron. Mater. 44, 658666 (2014).

49. Stauffer, D. Scaling theory of percolation clusters. Phys. Rep. 54, 174 (1979).50. Cai, L. et al. Super-stretchable , transparent carbon nanotube-based capacitive strain sensors for human motion detection. Sci. Rep. 3, 19 (2013).51. Hwang, B. et al. Transparent stretchable self-powered patchable sensor platform with ultrasensitive recognition of human activities. ACS Nano 9, 88018810 (2015).

52. Liu, H. et al. Electrically conductive thermoplastic elastomer nanocomposites at ultralow graphene loading levels for strain sensor applications. J. Mater. Chem. C 4, 157166 (2015).

53. Li, C. et al. Flexible CNT-array double helices strain sensor with high stretchability for motion capture. Sci. Rep. 5, 18 (2015).

54. Wu, X., Han, Y., Zhang, X. & Lu, C. Highly sensitive , stretchable , and wash-durable strain sensor based on ultrathin conductive layer @ polyurethane yarn for tiny motion monitoring. Appl. Mater. Interfaces 8, 99369945 (2016).

55. Chen, S., Wei, Y., Yuan, X., Lin, Y. & Liu, L. A highly stretchable strain sensor based on a graphene/silver nanoparticle synergic conductive network and a sandwich structure. J. Mater. Chem. C 4, 43044311 (2016).

56. Chae, S. H. & Lee, Y. H. Carbon nanotubes and graphene towards soft electronics. Nano Converg. 1, 1541 (2014).

57. Servant, A. et al. Graphene-based electroresponsive scaffolds as polymeric implants for on-demand drug delivery. Adv. Healthc. Mater. 3, 13341343 (2014).

58. Li, C., Adamcik, J. & Mezzenga, R. Biodegradable nanocomposites of amyloid brils and graphene with shape-memory and enzyme-sensing properties. Nat. Nanotechnol. 7, 421427 (2012).

59. Gong, J. P. Why are double network hydrogels so tough? Soft Matter 6, 25832590 (2010).

Acknowledgements

We thank Chaoxu Li, Tobias Keplinger and Jozef Adamcik for experimental assistance and discussions. We acknowledge internal funding from ETH Zrich, the Swiss National Science Foundation (grant 200020_146509 and Ambizione grant PZ00P2_148040) the Swiss Competence Center for Energy Research (SCCERCapacity Area A3: Minimization of energy demand) as well as support by the Center for Optical and Electron microscopy of ETH Zrich (ScopeM).

Author contributions

A.R.S. and R.M. conceived and supervised the study. Experiments were designed by all authors and conducted by H.L.F., S.B. and A.F.D., S.B. synthesized and characterized the modied graphene akes and conducted the scattering experiments. H.L.F. and A.F.D. designed and performed the magnetic patterning experiments and conductivity measurements together. H.L.F. prepared the composites and characterized their transparency and electrical conductivity. H.L.F. designed the gures; and H.L.F., R.L, A.R.S. and R.M. wrote the paper. H.L.F., S.B. and A.F.D. wrote the Supplementary Information. All authors analysed the data and discussed their implications and critically revised the manuscript at all stages.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Le Ferrand, H. et al. Magnetic assembly of transparent and conducting graphene-based functional composites. Nat. Commun. 7:12078doi: 10.1038/ncomms12078 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2016

NATURE COMMUNICATIONS | 7:12078 | DOI: 10.1038/ncomms12078 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9