Flexible cellulose acetate/graphene blueprints for vibrotactile actuator

Abstract

Tactile devices containing many actuators within are being sutured using electroactive polymers. This innovation forms the basis of hand-like tactile feedback in emerging smart robotic manipulation. Here, we introduce low power consuming modified reduced graphene oxide embedded cellulose acetate composite of high flexibility and conformability to surfaces made by simple and low cost synthesis methods, which thus points towards simple read-out electronics. The material performance was evaluated based on various actuation conditions in terms of electrical potential, bias voltage, temperature and frequency. The actuator vibrating at various frequencies with faster response time illustrates a range of haptic feedbacks to users, which can be used in braille display devices. Excellent repeatability of the haptic actuation process was also noticed.

---

1. Introduction

The interdisciplinary research fields, such as bio-inspired robotics, sensors, biomedical devices, touch-feedback haptic systems, and flexible soft electronics, exploit polymer based actuators driven by electrical stimuli to design and control virtual reality.¹ The word haptics refers to any form of nonverbal communication involving the application of a touch response in a user interface design to provide information to an end user. Very recently, haptics research has gained considerable importance as more physical and realistic virtual experience is created by the actuators embedded in the device. The tactile-feedback technology developed in mobile electronic devices provides touch sensation along with visual and auditory recognition and thus more realistic user experiences.^2–4 Such feedback devices using electrostatic force are also considered to be promising for touch based electronics.^5,6 Developing a full page refreshable braille tactile sensing display needs many small-diameter actuators with higher energy density to move the pins.^7,8 The actuators present in the commercially available displays require a significantly larger area than the braille pins themselves and so the displays are bulky, expensive and limited to two lines of braille characters. In this regard, designing a more compact and portable system capable of identifying the symbols and interfacing the blind community freely with computers has utmost importance.

Many types of braille displays used right now possess contractible multilayer stacked actuators,⁹ bistable electroactive polymers that use temperature control,¹⁰ dielectric buckling elastomer membranes^11,12 and polyvinylidene fluoride (PVDF) bending bimorph actuators coupled with hydraulic fluids.¹³ It is very necessary to build up eco-friendly tactile sensors of high flexibility and sensitivity at low-range voltage to apply in braille displays. Piezoceramic actuators are capable of producing vibrations at a wide frequency range from a small device,^14–16 but with difficulty in selectively stimulating mechanoreceptors as their vibrations are not enough except at their resonance frequencies. Instead of such materials, PVDF and copolymers have been studied for fabricating tactile sensors and actuators.^17,18 However, their melting nature and less availability promoted the invention of smart cellulose derivatives of light weight, eco-friendliness, low price, thermal stability, biocompatibility and biodegradability, very recently.^12,19–22 The cellulose acetate (CA) possesses potential compatibility, excellent optical and mechanical properties and good dielectric constant and is used to manufacture fibers, films, laminates, adhesives, coatings, plastic products and tactile actuators.^1,23,24 The electrically active nano-additives such as carbon nanotubes (CNTs) and graphene improve the mechanical properties of CA and enable the multi-functionality needed for electrical energy storage, sensing, and actuation.^1,19,25–28 Out of various nanocarbon additives, graphene is the most investigated due to its excellent electrical properties, mechanical strength, and flexibility.^29–32 The transistors, solar cells, sensors, batteries and supercapacitors based on flexible graphene render it suitable for a wide range of flexible electronic devices.^12,33,34 The outstanding capability of such materials to convert physical contacts into electric signals allows it to be utilized in high-performance flexible tactile sensors.^32,35,36

In this study, we present a new rolled CA/graphene composite that meets all of the vibrotactile actuator specifications with low voltage operation based on electrostatic potential. The fabrication process, performance evaluation and electrostatic behavior of this haptic actuator are discussed here on the basis of modification of the reduced graphene oxide (RGO) and the concentration of modified reduced graphene oxide (m-RGO). We analyze the morphologies and electrical and optical properties of CA/graphene composite by scanning electron microscopy (SEM), X-ray diffraction (XRD), semiconductor device analyzer and UV-visible spectrophotometer technologies as well.

2. Experimental section

Materials

Cellulose acetate powder (C₇₆H₁₁₄O₄₉), triethyl citrate (TEC, C₁₂H₂₀O₇), natural flake graphite, and other reagents, such as hexamethylene-1,6-diisocyanate (HMDI), dimethyl acetamide (DMAc), lithium chloride (LiCl), sulfuric acid (H₂SO₄), nitric acid (HNO₃), hydrochloric acid (HCl), potassium permanganate (KMnO₄) and 30% hydrogen peroxide (H₂O₂) solution, were procured from Sigma-Aldrich. 99.5% isopropyl alcohol and acetone were purchased from Daejung, South Korea.

Synthesis of m-RGO

The synthesis of GO was carried out by oxidizing graphite followed by the improved graphene oxide synthesis method.³⁷ The synthesized graphene oxide had a flake thickness of ∼1.5 nm and width ∼1.5 μm, as shown in Fig. S2.† The modification for RGOs was performed by dissolving the required amounts of GO and HMDI in DMF and sonicating for 10 min. The mixture was then refluxed for 48 h at 100 °C. The precipitate was filtered and washed several times with DMF to remove the extra methyl isocyanate.^21,38 The product was dried in a vacuum oven at 60 °C to obtain the functionalized GO nanosheets methyl isocyanate complex.

Preparation of CA/m-RGO composite

CA was dissolved in acetone at 7.5% (7.5 : 92.5) and mechanically stirred for about 3 hours at room temperature. Followed by this, TEC (plasticizer material) and homogeneous dispersions of 1 wt% m-RGO are dissolved one after the other in the solvent mixture by magnetic stirring. Subsequently, around 20 ml of solution was slowly poured in a silicon wafer and left in room temperature for drying. Finally, a flexible and transparent composite film typically 200 μm thick was obtained. In order to investigate the effect of the m-RGO sheets on the performance of the vibrotactile actuator, in addition to the CA/m-RGO composite film, a pristine CA film was also prepared. The electrodes for the study are made by spraying a thin layer of silver nanowires (AgNWs) on the side of the composite film. The spacers in between the two films allow the actuator to vibrate in plane perpendicular direction, and the finalized samples were tested on glass substrate at an ambient condition.

Characterization

Field emission SEM images of the sample films were taken with a JEOL JSM-6400F microscope to study the sample morphology. The samples were prepared by coating a platinum layer using an ion sputter (EMITECH, K575X). XRD patterns were checked with a thin film X-ray diffractometer using Cu-Kα target radiation at 40 kV and 50 mA, at a scanning rate of 0.015° min⁻¹. The diffraction angle was varied from 5° to 40°. Optical transmittance of the samples was studied using a UV-visible spectrophotometer. For this, the spectra of the films in the range of 200–700 nm wavelengths were recorded with a diode array (HP, 8452A). The variation in dielectric properties of the samples during 20 Hz–1 kHz frequency was monitored using an LCR meter (HP, 4284A). Measurements were done at room temperature. Setup for the vibrotactile actuator with a function generator (Agilent, 33220A), high voltage amplifier (Trek, 20/20), laser Doppler vibrometer (LDV; Ometron, VS-100), data acquisition system (B&K, PULSE) and LabVIEW software installed in a computer used to stimulate the actuator and measure vibration velocity is as shown in Fig. S1.† An element of the haptic actuator is made using film with patterned adhesive tape spacer, and then haptic actuator elements were arrayed and embedded in the device.

3. Results and discussion

Morphology

The morphology of the materials fabricated was observed with the help of SEM images. Fig. 1 shows the SEM pictures of the surface and cross sections of the CA and CA/m-RGO composite films. The morphology of CA consists of a smooth surface and cross-section, whereas in the CA/graphene composite case, a rough surface is seen due to large graphene flakes dispersed.

Fig. 1 SEM image of CA (a) surface and (c) cross section; CA/m-RGO (b) surface and (d) cross section.

The effect of filler nanosheets on the microstructure of the CA is very lucid from the pictures shown in Fig. 1b and d. The smooth fractured surface of pristine CA becomes rough due to the presence of m-RGO (Fig. 1d). Here the exfoliated m-RGO nanosheets are uniformly dispersed in the CA medium as evidenced by the CA/m-RGO, which can be attributed to the increased interaction between CA and m-RGO.

To further analyze the m-RGO dispersion in the polymer, the structural variation within the filler as well as the composites were investigated through XRD studies. As shown in the XRD pattern (Fig. 2a), the characteristic sharp peak of graphite at 2θ = 26.5° is more broadened (Fig. 2a) and shifted to 2θ = 5.3° in GO. This is due to the delamination of individual GO layers from graphite by the oxidation and sonication processes. This enhances the interlayer distance and causes a partial loss of regularity of the GO sheets. For GO, the characteristic peak is even less clear, indicating a wide range of interlayer spacing and a more substantial amount of exfoliation due to the surface modification.³⁹ The m-RGO, on the other hand, exhibits a peak corresponding to the d-spacing of 0.38 nm along the (002) orientation. This can be explained as the removal of oxygen functional groups causing a decrease in d-spacing. In the composite samples, the case is rather different (Fig. 2b). The neat CA peak obtained near 2θ = 21° is a crystalline halo attributed to its structure. For CA/m-RGO, the peak intensity marginally reduces though it remains at the same position at 2θ = 21°. This can be related to the crystallinity imparted to the CA matrix by the m-RGO addition. The m-RGOs possess a strong nucleating effect within the polymer matrix. In addition, the XRD spectrum of the samples shows no characteristic peak at 27° originating from the loss of regularity or exfoliation m-RGO in the CA, which also indicates the good dispersion of fillers in CA.

Fig. 2 XRD spectra of (a) graphite, GO and modified RGO and (b) CA and CA/m-RGO.

Optical transparency

Fig. 3 presents the UV spectra of the films showing the changes in transmittance values during the wavelengths 200–700 nm. Generally, the graphene sheets prepared by the reduction of m-GO contain many defects in chemical and crystal structure. Combining with the layer's overlapping in the film assemblage, the optical property of composite films is inferior to pristine graphene.²⁹ This is because the optical transmittance of graphene-based materials is highly dependent on their defects and the number of graphene layers.⁴⁰ The transmittance was recorded well over ∼80% throughout the entire visible wavelength for pristine CA film, whereas it reduced to around the 40% level for m-RGO added composite film, which can satisfy the requirements of certain semitransparent applications. Notably, the films possess very strong absorbance in the ultraviolet region, which would make them potential candidates for anti-UV applications.

Fig. 3 UV-visible spectra of CA and CA/m-RGO.

Dielectric properties

In order to investigate the effect of m-RGOs on the dielectric performance of CA film, frequency dependent capacitance as well dielectric loss was measured by a semiconductor device analyzer over a range of frequencies from 20 Hz to 10 kHz. Two different types of films were checked in this way, a pristine CA and a film with the CA/m-RGO composite, and the results are illustrated in Fig. 4. The dielectric constant (relative permittivity) of the composite film was calculated by eqn (1):where κ is the dielectric constant of the film, c is the capacitance, ε₀ the dielectric constant at vacuum, A and d are the area of the electrode and thickness of the film, respectively.

Fig. 4 Variation of dielectric constant of (a) CA and (b) CA/m-RGO with frequency.

At 150 Hz, the composite film showed a dielectric constant of 10.24 (0.032 dielectric loss), whereas the dielectric constant of CA was 5.66 with a dielectric loss of 0.027. Overall, the composite film of CA containing m-RGO exhibited around 80% increase in dielectric constant (0.037 dielectric loss) compared to the pristine CA (dielectric constant and dielectric loss were 5.54 and 0.033, respectively). Interfacial polarization at the interface between the m-RGO and the CA are attributed to the increase in dielectric constant of the composite film. Upon comparing the permittivity, CA/m-RGO composite possesses enhanced values compared with the neat due to the motion of free charge carriers causing interfacial polarization. The enhancement in the dielectric properties can be explained by the Maxwell–Wagner–Sillars process, in which the polymer–filler interfacial (such as donor–acceptor complexes) interaction plays the major role.⁴¹ The charge generated from metal coated electrodes will be trapped at the polymer filler interface upon applying the electric field and causes ultimate improvement in the dielectric property of the composite.^39,42

Vibrotactile performance

The response of the composite actuators under AC excitation of 200 and 500 V and input frequency 150 Hz is shown in Fig. 5a and b, respectively. The performance of the CA/m-RGO actuators increased significantly (three times higher) in comparison with those of the pure CA actuator. This is due to the contraction force produced by the electrostatic effect between the negative and positive electrodes to displace the top electrode in the downward direction. When a voltage is applied to the actuator, the sample film gets charged and then generates vibration. The motion of the actuator is concave, and the actuator performance was modulated by increasing the voltage level of the electric potential.

Fig. 5 Tactile performance in terms of velocity change vs. time at constant signal frequency 150 Hz and at voltage (a) 200 V and (b) 500 V.

There are two main methods for utilizing the electrostatic force: using a change in the electrostatic force, or charging two electrode layers separated by a dielectric material.^5,6,19 The high dielectric constant of the composite also causes high electromechanical response. The results indicate that the self-assembled composite can generate a considerably higher strain as well as elastic energy density compared with neat polymer film. Such enhancement can be partly attributed to exchange coupling that is very effective when the size of second phase particles approaches nanometer scale and there are large electric field fluctuations in the polymeric matrix due to the large contrast of constituent dielectric constants.¹ The improvement in permittivity with nanoparticle loading and according to the experiments, the actuator performance is directly influenced by the stiffness and dielectric constant of the samples. Apparently, the stiffness increases with nanoparticle loading at a faster rate than the dielectric permittivity does. Relative permittivity of CA/m-RGO composite film is two-fold larger than the pure CA film, which indicates that superior charge storing capacity in composite film is the dominant factor to create the electrostatic force. Due to the intensified electrostatic attraction force between chargeable films, performance is largely enhanced.

Frequency and voltage influence

Fig. 6a shows the contour plot of the measured vibration velocity of the actuators at various applied frequencies from 25 to 225 Hz (at a fixed voltage of 500 V). Note that the vibration velocity was measured at the center of the film actuator. The vibration performance decreased close to the spacers, and thus the points close to spacers provide a weaker tactile sensation in comparison to the center. It has also been reported that vibration at the low frequency range of 20–80 Hz is weakly felt by the finger, whereas vibration in the 150–250 Hz frequency range is most effective for information delivery through haptic feedback.¹⁹ Fig. 6b shows the responses of actuators of CA and CA/m-RGO composite films at 150 Hz with different voltage. Voltages ranging from 100 to 700 V were applied, and vibration velocity was measured at the touch layer of the actuators. The vibration velocity of the CA/m-RGO composite film actuator exponentially increased with voltage, whereas the performance of the CA film actuator increased linearly with applied voltage. Moreover, CA/m-RGO actuator was activated at low voltage, whereas CA films actuator performance was around noise level at 100 V. At 700 V, the maximum vibration velocity of the CA/m-RGO actuator reaches 20 mm s⁻¹, whereas it was around 4 mm s⁻¹ for pure CA film samples. All the above-mentioned performance measurement conditions clearly show the superior performance of the CA/m-RGO actuator over the pure CA film actuator.

Fig. 6 Velocity vs. frequency at constant (a) voltage 500 V and (b) signal frequency 150 Hz for CA and its composite film.

Biased voltage

Fig. 7 shows the bias voltage effect on the velocity of the CA/m-RGO actuator under a consistent sine wave signal of low level voltage at 150 Hz. With an increase in the bias voltage, the output of the CA/m-RGO actuator increases. It suggests that bias voltage can be another important factor for a further improvement of the actuator performance. Higher fixed charge loading with bias voltage can lead to denser surface charge, which produces higher electrostatic attraction force between electrodes of the actuator. Moreover, biased voltage further reduced the activation voltage of the CA/m-RGO actuator to 50 V.

Fig. 7 Bias voltage effect on CA/m-RGO actuation performance (a) velocity vs. time and (b) velocity vs. bias voltage plots.

Temperature influence

The performance of actuators prepared by CA and CA/m-RGO composite films at 150 Hz applied frequency by varying the temperature is shown in Fig. 8. As the temperature increases, the performance curve of the composite film actuator shows a steep decrease and fall around the level of the pure CA film actuator. Moreover, the pure CA film actuator exhibits constant performance during the temperature range 25–60 °C. It may be given by the fact that the relaxation of chains, which are not in direct contact with the filler surface, is considerably more temperature-dependent than the retarded relaxation motion of chains in the interphase region interacting directly with the filler surface. In other words, increasing temperature causes considerably faster softening of the bulk matrix in comparison with immobilized polymer chains (bound to the filler). Thus, an increase of the temperature of the composite actuator film causes rupture in the weak physical polymer–filler bonds (chain desorption) arising at high strain amplitudes and results in decreased performance.³⁸

Fig. 8 Temperature influence on tactile performance of CA and CA/m-RGO actuators.

Proof-of-concept for braille display

One way of solving the touch feeling issue is to create haptic actuators that can generate various button sensations via the vibrotactile principle. There have been many studies to imitate button sensation using responsive actuators because tactile actuators can be easily designed in small dimensions. A hardware prototype is presented that has been built to analyze the transmission characteristics of the braille board using actuation. The braille cell prototype was designed and fabricated using CA/graphene film as actuators. The prototype was used next to examine possible actuator efficiency in force-feedback control with the designed actuator, as shown in Fig. 9.

Fig. 9 Schematic representation of braille cell using vibrotactile actuator.

The mimicking button sensation with vibration has a wide-frequency bandwidth (25–225 Hz), and the response time of the vibrotactile actuator will be smaller than the temporal resolution for the haptic sensor (about 20 ms). Interestingly, it is worth noting that for now, flexible touch displays are used in curved and hard surfaces, in which conventional tactile-feedback devices may be still useful. We observed that the proposed vibrotactile actuator is small enough to be embedded in handheld devices with fast response time and creates a vibrotactile force large enough to stimulate all mechanoreceptors. As shown in Fig. S3,† actuator performance has limitations depending on the type of wave function, and sinusoidal form shows better performance than ramp and square type. Moreover, the enhancement of mechanical properties is also obtained with the presence of m-RGO (Fig. S4†). We also analyzed the repeatability of the actuator by cyclic on and off voltage, and it was found to have good reproducibility. The technology promises to bring new forms of tactile experience that will expand touch interaction with equipment and make them feel more natural while improving productivity, eco-friendliness and compatibility. The present advanced electronic devices adopting haptic technology rapidly spreads from a handful of consumer applications to a much more extensive scope of industrial, commercial, automotive, medical and different frameworks.

4. Conclusion

An actuator based on cellulose derived graphene composite was developed and applied for the preferable execution of a braille tactile display. The proposed system exhibits controlled vibrotactile performance with respect to vibration amplitude, bias voltage and frequency. At the point when the film vibrator was working at its lowest voltage (50 V) and at a frequency of 150 Hz, enough haptic sensation was observed at around one μm of vibration amplitude. This device shows excellent actuator performance. However, its performance degrades with time and high temperature conditions. A proof-of-concept was explicated with lightweight, flexible, cheap, compact and expeditious time and furthermore gives touch-coordinate particular replication of braille display.

Acknowledgements

This work was supported by National Research Foundation (NRF-2013M3C1A3059586), Republic of Korea.

References

1. M. Mohiuddin, H.-C. Kim, S. Y. Kim and J. Kim, Proc. SPIE, 2014, 9060, 906017.
2. L. Thurfjell, J. McLaughlin, J. Mattsson and P. Lammertse, Ind. Robot, 2002, 29, 210–215.
3. S. D. Laycock and A. M. Day, Comput. Graph. Forum, 2003, 22, 117–132.
4. K. J. Kuchenbecker, J. Fiene and G. Niemeyer, IEEE Trans. Visual. Comput. Graph., 2006, 12, 219–230.
5. O. Bau, I. Poupyrev, A. Israr and C. Harrison, Proc. of the 23nd annual ACM symp. On UIST, 2010, 10, 283–292.
6. M.-C. Choi, Y. Kim and C.-S. Ha, Prog. Polym. Sci., 2008, 33, 581–630.
7. A. N. Sokolov, B. C.-K. Tee, C. J. Bettinger, J. B.-H. Tok and Z. Bao, Acc. Chem. Res., 2011, 45, 361–371.
8. T. Someya, Y. Kato, T. Sekitani, S. Iba, Y. Noguchi, Y. Murase, H. Kawaguchi and T. Sakurai, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 12321–12325.
9. G. Kovacs and L. During, Proc. SPIE, 2009, 7287, 72870A.
10. Y. Hui-Jian, Y. Li, S. Wen-Zhu, S. Song-Bai and Z. Liang, RSC Adv., 2013, 3, 23730–23736.
11. F. Carpi, G. Frediani and D. D. Rossi, Proc. SPIE, 2010, 7642, 76420E.
12. K. Ren, L. Sheng, L. Minren, W. Yong and Q. M. Zhang, Sens. Actuators, A, 2008, 143, 335–342.
13. N. D. Spigna, P. Chakraborti, D. Winick, P. Yang, T. Ghosh and P. Franzon, Proc. SPIE, 2010, 7642, 76420A.
14. A. B. Dalton, S. Collins, J. Razal, E. Munoz, V. H. Ebron, B. G. Kim, J. N. Coleman, J. P. Ferraris and R. H. Baughman, J. Mater. Chem., 2004, 14, 1–3.
15. R. J. Wood, E. Steltz and R. S. Fearing, Sens. Actuators, A, 2005, 119, 476–488.
16. D.-H. Kim, B. Kim and H. Kang, Microsyst. Technol., 2004, 10, 275–280.
17. M. Wagner, A. Roosen, H. Oostra, R. Hoppener and M. D. Moya, J. Electroceram., 2005, 14, 231–238.
18. A. V. Bune, C. Zhu, S. Ducharme, L. M. Blinov, V. M. Fridkin, S. P. Palto, N. N. Petukhova and S. G. Yudin, J. Appl. Phys., 1999, 85, 7869–7873.
19. D. Ponnamma, K. K. Sadasivuni, M. Strankowski, Q. Guo and S. Thomas, Soft Matter, 2013, 9, 10343–10353.
20. A. Ambrosy and K. Holdik, J. Phys. E: Sci. Instrum., 1984, 17, 856–859.
21. A. Kafy, K. K. Sadasivuni, H.-C. Kim, A. Akther and J. Kim, Phys. Chem. Chem. Phys., 2015, 17, 5923–5931.
22. K. K. Sadasivuni, M. Yadav, X. Gao, S. Mun and J. Kim, Proc. SPIE, 2014, 9060, 906016.
23. H. S. Barud, A. M. D. Araújo, D. B. Santos, R. M. N. D. Assunção, C. S. Meireles, D. A. Cerqueira, G. R. Filho, C. A. Ribeiro, Y. Messaddeq and S. J. L. Ribeiro, Thermochim. Acta, 2008, 471, 61–69.
24. D.-G. Yu, J.-H. Yu, L. Chen, G. R. Williams and X. Wang, Carbohydr. Polym., 2012, 90, 1016–1023.
25. M. K. Shin, B. Lee, S. H. Kim, J. A. Lee, G. M. Spinks, S. Gambhir, G. G. Wallace, M. E. Kozlov, R. H. Baughman and S. J. Kim, Nat. Commun., 2012, 3, 650.
26. A. B. Dalton, S. Collins, E. Muñoz, J. M. Razal, V. H. Ebron, J. P. Ferraris, J. N. Coleman, B. G. Kim and R. H. Baughman, Nature, 2003, 423, 703.
27. N. Lachman, C. Bartholome, P. Miaudet, M. Maugey, P. Poulin and H. D. Wagner, J. Phys. Chem. C, 2009, 113, 4751–4754.
28. M. Gopiraman, K. Fujimori, K. Zeeshan, B. S. Kim and I. S. Kim, eXPRESS Polym. Lett., 2013, 7, 554–563.
29. K. K. Sadasivuni, D. Ponnamma, J. Kim and S. Thomas, Graphene-Based Polymer Nanocomposites in Electronics, Springer publisher, Switzerland, 2015, p. 67.
30. K. K. Sadasivuni, D. Ponnamma, S. Thomas and Y. Grohens, Prog. Polym. Sci., 2014, 39, 749–780.
31. D. Ponnamma, K. K. Sadasivuni, M. Strankowski, P. Moldenaers, S. Thomas and Y. Grohens, RSC Adv., 2013, 3, 16068–16079.
32. T. Yamada, Y. Hayamizu, Y. Yamamoto, Y. Yomogida, A. Izadi-Najafabadi, D. N. Futaba and K. Hata, Nat. Nanotechnol., 2011, 6, 296–301.
33. U. Kim, J. Kang, C. Lee, H. Y. Kwon, S. Hwang, H. Moon, J. C. Koo, J.-D. Nam, B. H. Hong, J.-B. Choi and H. R. Choi, Nanotechnology, 2013, 24, 145501.
34. K. K. Sadasivuni, A. Kafy, L. Zhai, H.-U. Ko, S. Mun and J. Kim, Small, 2015, 11, 994–1002.
35. N. O. Weiss, H. Zhou, L. Liao, Y. Liu, S. Jiang, Y. Huang and X. Duan, Adv. Mater., 2012, 24, 5782–5825.
36. X. Cao, D. Qi, S. Yin, J. Bu, F. Li, C. F. Goh, S. Zhang and X. Chen, Adv. Mater., 2013, 25, 2957–2962.
37. D. C. Marcano, D. V. Kosynkin, J. M. Berlin, A. Sinitskii, Z. Sun, A. Slesarev, L. B. Alemany, W. Lu and J. M. Tour, ACS Nano, 2010, 4, 4806–4814.
38. K. K. Sadasivuni, D. Ponnamma, B. Kumar, M. Strankowskie, R. Cardinaels, P. Moldenaers, S. Thomas and Y. Grohens, Compos. Sci. Technol., 2014, 104, 18–25.
39. S. Littlejohn, A. Nogaret, G. Prentice and G. Pantos, Adv. Funct. Mater., 2013, 23, 5398–5402.
40. S. Pang, Y. Hernandez, X. Feng and K. Mullen, Adv. Mater., 2011, 23, 2779–2795.
41. J.-K. Yuan, S.-H. Yao, Z.-M. Dang, A. Sylvestre, M. Genestoux and J. Bai, J. Phys. Chem. C, 2011, 115, 5515–5521.
42. K. K. Sadasivuni, M. Castro, A. Saiter, L. Delbreilh, J. F. Feller, S. Thomas and Y. Grohens, Mater. Lett., 2013, 96, 109–112.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra03043a

---