On the nucleation of graphene by chemical vapor deposition

Abstract

We demonstrate that homogeneous single-layer graphene can be grown by simply annealing crystalline Cu(111)/c-plane sapphire (α-Al₂O₃) at 900 and 1000 °C without additional carbon supply. The resulting graphene film shows a high carrier mobility of 1210 cm² V⁻¹ s⁻¹. However, the annealing at a lower temperature of 800 °C gives an amorphous carbon film. Further investigations indicate that graphitization of amorphous carbon and/or adsorbed carbon atoms during chemical vapour deposition (CVD) is not simply a consequence of carbon supersaturation, it is also affected by CVD temperature, and crystallographic plane of the underlying metal, which are essentially correlated to the energy barrier of nucleation. Our results provide the direct experimental evidence to elucidate the influencing factors of graphitization of amorphous carbon, and contribute fundamental insight into the nucleation and growth of graphene to improve its quality for applications.

---

1. Introduction

Graphene is an ideal two dimensional material for many promising applications, such as transistors,¹ transparent electrodes,² liquid crystal devices,³ and solar cells.⁴ The applicable preparations of single- and few-layer graphene sheets with low cost are expected. Among several most prominent methods, chemical vapor deposition (CVD) on transition metals with a supplied C source, mainly CH₄ gas, can accomplish the easily transferable and scalable graphene films, mostly by using polycrystalline Ni films,^5,6Ni foils,^7,8Cu films,^9,10 and Cu foils.^11,12

Due to low C solubility, the Cu metal enables the growth of uniform single-layer graphene facilely, and is thought as a potential metal for the large-scale production of CVD-derived graphene. Recently, several groups have reported excellent device characteristics such as mobilities of up to 7350 cm² V⁻¹ s⁻¹ and large area growth (up to 30 inch);¹³ however, a large amount of domain boundaries and a relatively small domain size in the formed graphene film still limit its properties. It has been suggested that graphene nucleates at defects on the Cu surface imperfections and/or grain boundary edges, which eventually coalesce to form a continuous layer of polycrystalline graphene with numerous domain boundaries.^14–16 It is thus observed that suppressing the density of the nucleation site can effectively increase the domain size. Several reports have confirmed this. For example, Liet al. found that modification of growth parameters can reduce the density of graphene nuclei for increasing the domain size of graphene;¹⁴ additionally, the use of single crystalline Cu(111) is applicable to reduce the concentration of nucleation sites (surface imperfections, grain boundary edges), in turn decreasing the concentration of domain boundaries.¹⁷ These results indicate that the control over nucleation of an individual graphene domain affords potential to remarkably improve the graphene quality; therefore, more investigation on graphene nucleation is crucial.

It is well-acknowledged that formation of graphene nuclei is a consequence of local supersaturation of C atoms in a metal.^11,14 However, very recently, graphitization of amorphous carbon by direct metal-induced crystallization was observed in situ;¹⁸ also, the Cu metal was found to dominate the shape and orientation of nucleation lands of graphene.¹⁵ These results remind us of a long-term question of whether the nature of the metal affects the formation of graphene nuclei. To distinguish graphitization of amorphous carbon by direct metal-induced crystallization from the mechanism of graphene nucleation attributed to supersaturation of C atoms, a critical amount of C source for graphene growth is necessary. It is recognized that such a trace amount of carbon inevitably exists on the metal surface/bulk,¹⁹ and can form graphene by thermal annealing of this metal.^20,21 In this article, we have grown uniform single-layer graphene by annealing the Cu(111) film deposited on c-plane sapphire (α-Al₂O₃) without additional carbon supply, and systematically investigate the influencing factors of graphitization of amorphous carbon into graphene, which has not been reported thus far.

2. Results and discussion

2.1 Effect of annealing temperature

When annealing the Cu/c-plane Al₂O₃ at 800 °C and after transfer, some discrete films with low coverage were observed on the SiO₂/Si substrate, as shown in Fig. 1a. A typical Raman spectrum (Fig. 1d) displays broad D and G bands without a representative peak of graphite at ∼2690 cm⁻¹ (2D band), indicating the formation of amorphous carbon.²³ When the annealing temperatures were elevated to 900 and 1000 °C, we found the growth of continuous and uniform graphene films, as seen in Fig. 1b and c. The Raman intensity ratios of G to 2D bands (I_G/I_2D ratio) are about 0.45–0.55 (Fig. 1d). The full widths at half maximum (FWHM) of 2D bands are ∼35 cm⁻¹, and the peak profile of 2D bands can be fitted by single Lorentzian. These data confirm the formation of single-layer graphene.^9–11 To further evaluate the homogeneity of as-grown single-layer graphene film, we mapped the Raman intensities of G (∼1585 cm⁻¹) and 2D bands (∼2685 cm⁻¹), as shown in Fig. 1e and f. In the measured area of 20 μm × 20 μm, I_G/I_2D ratios show the values of ∼0.45, denoting a large fraction of single-layer graphene. Fig. 2a (inset, left) also shows an image of uniform single-layer, large area (15 mm × 15 mm) graphene film which was transferred onto a target SiO₂/Si substrate. This suggests that simple annealing without additional carbon supply enables the large-scale growth of graphene.

Fig. 1 Optical micrographs of transferred graphene films from the Cu/c-plane Al₂O₃ substrates at the annealing temperatures of (a) 800, (b) 900, and (c) 1000 °C. (d) shows the typical Raman spectra of the graphene films measured with 514.5 nm excitation. (e) and (f) are the Raman mapping images of G and 2D band intensities measured for the graphene film (c).

Fig. 2 (a) Optical transmittance of single-layer graphene film transferred on a quartz substrate. The graphene was grown on Cu/Al₂O₃ by annealing at 1000 °C. The light transmission through a hole with 1.8 mm diameter was measured. Inset is the photograph of the transferred graphene films on SiO₂/Si (left) and quartz (right). (b) Transfer curve of graphene FET fabricated on SiO₂/Si (channel width and length are 50 μm and 7.5 μm, respectively).

To further evaluate the quality of graphene film, we measured the light transmission and transport characteristics. Fig. 2a (inset, right) shows the transferred graphene film on quartz, and the light transmission profile exhibits a value of 97.3% at 550 nm, confirming the single-layer graphene,^24,25 which is consistent with the above Raman data. Fig. 2b is the transport curve of a field-effect transistor (FET) fabricated with the graphene transferred onto the SiO₂/Si substrate. The hole- and electron-mobilities were calculated to be 1210 and 1000 cm² V⁻¹ s⁻¹, respectively. These mobilities are relatively high among the CVD-derived graphene-based devices using the Cu metal reported so far.^6,11 It is seen that the Dirac point is shifted to positive gate voltage, showing p-type behavior; this is probably due to remaining FeCl₃ introduced by the transfer process or physisorption of small molecules, such as H₂O or O₂.²⁶ The mobility of the graphene-based FET device can be improved with a proper choice of gate dielectric¹¹ and optimization of the fabrication process.

2.2 Effect of the crystallographic nature of the Cu substrate

We performed X-ray diffraction (XRD) to analyze the crystallinity of different Cu metals after 900 °C annealing. As seen in Fig. 3a and b, Cu film on the c-plane Al₂O₃ is highly crystalline with the fcc(111) plane normal to the surface, while that on the SiO₂/Si has a polycrystalline structure which is mainly composed of fcc(111), associated with a small fraction of the fcc(100) plane. The Cu foil (I) shows a strong fcc(110) diffraction band and two weak peaks corresponding to fcc(111) and fcc(100) (Fig. 3c); the Cu foil (II) shows a diffraction peak of highly crystalline fcc(100) (Fig. 3d).

Fig. 3 XRD profiles of Cu films/foils after annealing at 900 °C with H₂/Ar gas. (a) Cu/c-plane Al₂O₃, (b) Cu/SiO₂/Si, (c) Cu foil (I), and (d) Cu foil (II).

Subsequently, we annealed Cu/SiO₂/Si and two kinds of Cu foils at 900 °C, as seen in Fig. 4a, with the same H₂/Ar gas flow as that used for Cu/Al₂O₃. After annealing Cu/SiO₂/Si, optical microscopy and Raman spectra indicate the formation of single-layer graphene with some few-layer flakes, but the D band is relatively stronger than that for Cu/Al₂O₃. This reveals that the underlying substrate affects the quality of Cu metal and thus the resulting graphene. Differently, the two kinds of Cu foils gave the large-area amorphous carbon films (Fig. 4b and c), and even at a higher annealing temperature of 1000 °C (not shown). Hence, similarly to the annealing temperature, different Cu substrates affect the formation of C products (amorphous carbon or graphite), which may be attributed to the crystallographic nature of the metal, as will be discussed later.

Fig. 4 Raman spectra of transferred graphene films with 900 °C annealing of (a) Cu/SiO₂/Si, (b) Cu foil (I), and (c) Cu foil (II). Insets show the corresponding optical micrographs.

2.3 Effect of carbon supply

To compare growth of graphene by CH₄-free annealing with CH₄-CVD, we introduced the additional CH₄ for 10 min to perform conventional CH₄-CVD on the above Cu films/foils at 800–1000 °C with the same gas flow of CH₄/H₂/Ar = 20/20/400 sccm. In contrast to the carbon supply-free cases shown in Fig. 1a and 4b and c, graphene films can be grown at 800 °C on the Cu/Al₂O₃, and at 900 °C on the two Cu foils with CH₄-CVD (Fig. 5). These results confirm that only if sufficient carbon source is supplied, graphene can be formed, irrespective of growth temperature (800–1000 °C) and nature of the metal, suggesting that the carbon supply is an important influencing factor of graphene nucleation. This may account for why supersaturation of C atoms in the metal has been solely emphasized in the formation of graphene nuclei so far. In agreement with our findings, Liet al. reported that graphene nucleation can happen only if the flow rate and partial pressure of CH₄ are larger than critical values.¹⁴ It should be explained that single layer graphene can be obtained on Cu/SiO₂/Si or Cu foils through optimization of the growth conditions, as evidenced in previous work.^9–12

Fig. 5 (a), (c), (e), and (g) are the optical micrographs of transferred graphene films with CH₄-CVD from different Cu films/foils. (b), (d), (f), and (h) are the typical Raman spectra measured at the marked positions of the graphene films (a), (c), (e), and (g), respectively.

2.4 Discussion on influencing factors of graphitization of amorphous carbon into graphene

All the above evoke to elucidate under what conditions graphitization (or crystallization) of carbon can initialize. It has been established that the graphene nuclei stem from crystallization of C atoms and amorphous carbon forms in the absence of graphene nuclei. In crystallography, the change in Gibbs free energy as a function of number of atoms in the crystalline phase, G(N) = E(N) − Δμ × N, dominates the behavior of nucleation and growth of crystals; where, Δμ is the chemical potential difference between this crystalline phase and atom source, E is the total energy of C atoms, N is the number of C atoms.²⁷ The nucleus size and nucleation barrier (N*, G*) are defined as the maximum of the G(N) curve. Gao et al. calculated that the nucleation barrier G* is lowered with the increase of Δμ.²⁸

Combining the above XRD data with the carbon products, as summarized in Table 1, one can see that high annealing temperature (Fig. 1) and the sufficient carbon supply (Fig. 5) favor the graphitization of C atoms. Following the above definition, we suggest that higher annealing temperatures of 900 and 1000 °C for the Cu/c-sapphire can increase the chemical potential (Δμ) of carbon atoms and thus decrease the nucleation barrier (E), as computed theoretically,²⁹ which results in the facile formation of the graphene nuclei. Similarly, Liu et al. proposed that sufficient temperature is necessary to provide the minimal essential energy for converting the segregated carbon atoms into the sp² graphitic state.³⁰ Our conclusion is based on an assumption that all the C atoms originate from the Cu surface and the amounts of C atoms are almost the same at the three annealing temperatures, because our Cu target is of high purity and those C atoms that have diffused out of the Cu bulk can be ignored, as shown in Fig. S1 (ESI†). This result also explains why annealing of Cu/Al₂O₃ at 800 °C gave the amorphous carbon film, and annealing at 900 and 1000 °C led to the formation of graphene, as shown above in Fig. 1. On the other hand, sufficient carbon supply can promote the local supersaturation of C atoms, enhancing the graphene nucleation, as is well known.^6,11 This is also responsible for the graphene growth with CH₄-CVD at 800 °C, in contrast to the case of CH₄-free annealing. In fact, this mechanism can also be interpreted by the reduction of nucleation barrier mentioned above, because high carbon coverage increases the chemical potential of carbon atoms.³¹

Table 1 Carbon products formed on different Cu substrates and at different temperatures

	Cu/c-plane Al2O3		Cu/SiO2/Si	Cu foil (I)	Cu foil (II)
a Amorphous carbon. b Single layer graphene. c Few layer graphene.
Temp./°C	800	900	900	900	900
XRD diffractions	(111)	(111)	(111), (200)	(220), (200)	(200)
Without CH4	a-C ^a	SLG ^b	SLG & FLG^c	a-C	a-C
With CH4	FLG	SLG	FLG	FLG	FLG

---

Next, we discuss the effect of Cu metal substrate. Above results show that the annealing of crystalline Cu/c-sapphire and polycrystalline Cu/SiO₂/Si at 900 °C gave graphene films, but the annealing of crystalline Cu foil(II) led to amorphous carbon. Thus, single- or poly-crystallinity of the Cu metal is not the predominant factor of graphene nucleation. It is also seen in Table 1 and Fig. 3 that the annealing of Cu films on c-plane Al₂O₃ and SiO₂/Si, which are mainly composed of Cu(111) crystal planes, both gave graphene; while the annealing of Cu foils (I) and (II) which are mainly composed of Cu(110) and/or Cu(100) both resulted in amorphous carbon films at the same annealing temperature of 900 °C, and even at a higher temperature of 1000 °C which favors the formation of graphene nuclei as demonstrated above. These results lead us to speculate that the Cu(111) crystal plane prefers the graphene nucleation than Cu(100) and (110) crystal planes when the C supply is very low. This is probably because that the triangular Cu(111) plane is prone to have more similarities in crystallographic geometry as the hexagonal graphene lattice than other crystal planes, thus reducing the nucleation barrier, since the lattice mismatch between graphene and underlying metal will cause an additional energy cost (E is increased).^29,32 This result is also supported by the previous report that the additional energy cost due to the lattice mismatch can result in the destabilization of graphene nuclei.²⁹ Although there has been some concern that graphene may preferentially nucleate and grow on specific Cu crystallographic surfaces,³³ our results provide the first CVD-derived experimental evidence. Such similarity of the crystal structure and low lattice disregistry may be a common influencing parameter of nucleation and crystallization, as is the case of metal crystallization.³⁴

The carbon source is another important issue to be addressed. Considering high purities of sputtered Cu film (secondary ion-microprobe mass spectrometry (SIMS) analysis indicated the minimum level of C concentration inside the sputtered Cu film, as seen in ESI†, Fig. S1), Cu foil, and H₂–Ar gases, we speculate that the C source comes from inevitable C deposited on the surface of the metal and/or substrate, and/or C desorption from the inner wall of the quartz tube. To specify the unintentional carbon source, we labeled the introduced carbon source by ¹³C. Fig. S2 (ESI†) displays the Raman spectra of the transferred graphene film after introducing ¹³CH₄. Separated ¹²C- and ¹³C-graphene domains appeared, which can be explained by the surface adsorption mechanism proposed by Liet al.³⁵ Hence, the unintentional ¹²C atoms have already existed prior to the introduced ¹³C. Otherwise, the graphene should have ¹²⁺¹³C domains. Although further analysis is necessary to determine the C source and quantify the level of C contamination, present data can logically serve to decide whether graphene or amorphous carbon is formed and further to elucidate the influencing factors of graphitization of C. Additionally, they are not expected to discern the remarkable function of as-grown single-layer graphene. The different levels of C contamination were claimed to influence the as-grown graphene in C-containing Ni or Co films;^30,36 however, this remarkable effect is not expected for the as-grown single-layer graphene on Cu metal.

3. Conclusions

Continuous and uniform single-layer graphene is obtained by simply annealing the C-contaminated sputtered Cu film deposited on c-plane Al₂O₃. We investigated the influencing factors of graphitization of C into graphene, and clarified that with a quite small amount of C source, a Cu(111) crystallographic plane and a relatively high temperature are beneficial to nucleation and growth of graphene; whereas, with a sufficient C supply, the C supersaturation appears predominant for the formation of graphene, irrespective of the temperature and crystallographic nature. Although more pronounced studies are required, our results are applicable to present the experimental evidence that not only sufficient C supply, but also the nature of the metal imposes an effect at the nucleation stage of graphene, which extends deeper understanding of graphene growth and favors further improvement of the quality of CVD-derived graphene.

4. Experimental

500 nm-thick Cu films were deposited onto c-plane Al₂O₃ and SiO₂(300 m)/Si substrates by magnetron sputtering using a 99.999% pure Cu target. For comparison, two different Cu foils, Cu foil (I) (Nilaco, 25 μm thickness, 99.99% purity) and Cu foil (II) (Tanaka Kikinzoku Kogyo K. K., 30 μm thickness, 99.99% purity), were also studied. To clean a quartz tube used for annealing, the tube was flashed for 2 h and pre-baked for 1 h at 1000 °C with Ar flow. After cooling the tube to room temperature, the as-sputtered Cu film (or foil) was loaded and annealed for 1 h with a gas flow of 20 sccm H₂ (99.99999% purity) and 400 sccm Ar (99.999% purity) under ambient pressure at 800–1000 °C, followed by the rapid cooling to room temperature. Then, the as-formed film was transferred from the Cu surface onto a SiO₂/Si wafer using PMMA spin coating and FeCl₃/HCl solution.²²Raman spectra and mapping images of graphene films were measured with JASCO NRS-2100 using 514.5 nm excitation wavelength. The crystallinity of Cu films/foils was evaluated by an X-ray diffraction (XRD) (RIGAKU RINT 2500). Optical transmittance of the as-transferred graphene film was measured with a spectrophotometer (Jasco V-570). The back-gated FET measurement was carried out in a vacuum (∼5 × 10⁻⁴ Pa) with a semiconductor analyzer (Agilent, B1500A) using a probe station at room temperature.

Acknowledgements

This work was supported by the JSPS Funding Program for Next-Generation World-Leading Researchers (NEXT program) and PRESTO-JST. We acknowledge K. Tanaka of Kyushu Univ. for the SIMS measurement.

Notes and references

1. T. J. Echtermeyer, M. C. Lemme, M. Baus, B. N. Szafranek, A. K. Geim and H. Kurz, IEEE Electron Device Lett., 2008, 29, 952–954.
2. K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, J. H. Ahn, P. Kim, J. Y. Choi and B. H. Hong, Nature, 2009, 457, 706–710.
3. P. Blake, P. D. Brimicombe, R. D. Nair, T. J. Booth, D. Jiang, F. Schedin, L. A. Ponomarenko, S. V. Morozov, H. F. Gleeson, E. W. Hill, A. K. Geim and K. S. Novoselov, Nano Lett., 2008, 8, 1704–1708.
4. J. B. Wu, H. A. Becerril, Z. N. Bao, Z. F. Liu, Y. S. Chen and P. Peumans, Appl. Phys. Lett., 2010, 92, 263302.
5. Y. Zhang, L. Gomez, F. N. Ishikawa, A. Madaria, K. Ryu, C. Wang, A. Badmaev and C. W. Zhou, J. Phys. Chem. Lett., 2010, 1, 3101–3107.
6. Y. Lee, S. Bae, H. Jang, S. Jang, S. Zhu, S. H. Sim, Y. Song, B. H. Hong and J. H. Ahn, Nano Lett., 2010, 10, 490–493.
7. H. Cao, Q. Yu, R. Colby, D. Pandey, C. S. Park, J. Lian, D. Zemlyanov, I. Childres, V. Drachev, E. A. Stach, M. Hussain, H. Li, S. S. Pei and Y. P. Chen, J. Appl. Phys., 2010, 107, 044310.
8. Y. Miyata, K. Kamon, K. Ohashi, R. Kitaura, M. Yoshimura and H. Shinohara, Appl. Phys. Lett., 2010, 96, 263105.
9. M. Levendorf, C. Ruiz-Vargas, S. Garg and J. Park, Nano Lett., 2009, 9, 4479–4483.
10. Y. H. Lee and J. H. Lee, Appl. Phys. Lett., 2010, 96, 083101.
11. X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung, B. Tutuc, S. K. Banerjee, L. Colombo and R. S. Ruoff, Science, 2009, 324, 1312–1314.
12. S. Bhaviripudi, X. Jia, M. S. Dresselhaus and J. Kong, Nano Lett., 2010, 10, 4128–4133.
13. S. Bae, H. Kim, Y. Lee, X. Xu, J. S. Park, Y. Zheng, J. Balakrishnan, T. Lei, H. R. Kim, Y. I. Song, Y. J. Kim, K. S. Kim, B. Ozyilmaz, J. H. Ahn, B. H. Hong and S. Iijima, Nat. Nanotechnol., 2010, 5, 574–578.
14. X. Li, C. W. Magnuson, A. Venugopal, J. An, J. W. Suk, B. Han, M. Borysiak, W. Cai, A. Velamakanni, Y. Zhu, L. Fu, E. M. Vogel, E. Voelkl, L. Colombo and R. S. Ruoff, Nano Lett., 2010, 10, 4328–4334.
15. J. M. Wofford, S. Nie, K. F. McCarty, N. C. Bartelt and O. D. Dubon, Nano Lett., 2010, 10, 4890–4896.
16. P. Y. Huang, C. S. Ruiz-Vargas, A. M. van der Zande, W. S. Whitney, M. P. Levendorf, J. W. Kevek, S. Garg, J. S. Alden, C. J. Hustedt, Y. Zhu, J. Park, P. L. McEuen and D. A. Muller, Nature, 2011, 469, 389–392.
17. (a) L. Gao, J. F. Guest and N. P. Guisinger, Nano Lett., 2010, 10, 3512–3516; (b) K. M. Reddy, A. D. Gledhill, C. H. Chen, J. M. Drexler and N. P. Padturea, Appl. Phys. Lett., 2011, 98, 113117.
18. K. L. Saenger, J. C. Tsang, A. A. Bol, J. O. Chu, A. Grill and C. Lavoie, Appl. Phys. Lett., 2010, 96, 153105.
19. J. W. May, Surf. Sci., 1969, 17, 267–270.
20. A. J. Pollard, S. N. Sabki, C. R. Staddon, L. M. A. Perdigao, C. H. Hsu, J. M. Garfitt, S. Gangopadhyay, H. F. Gleeson, A. K. Geim and P. H Beton, J. Phys. Chem. C, 2009, 113, 16565–16567.
21. R. G. Musket, W. McLean, C. A. Colmenares, D. M. Makowiecki and W. J. Siekhaus, Appl. Surf. Sci., 1982, 10, 143–207.
22. H. Ago, I. Tanaka, C. M. Orofeo, M. Tsuji and K. Ikeda, Small, 2010, 6, 1226–1233.
23. H. J. Scheibe, D. Drescher and P. Alers, Fresenius' J. Anal. Chem., 1995, 353, 695–697.
24. X. Li, Y. Zhu, W. Cai, M. Borysiak, B. Han, D. Chen, R. D. Piner, L. Colombo and R. S. Ruoff, Nano Lett., 2009, 9, 4359–4363.
25. R. R. Nair, P. Blake, A. N. Grigorenko, K. S. Novoselov, T. J. Booth, T. Stauber, N. M. R. Peres and A. K. Geim, Science, 2008, 320, 1308.
26. Z. Sun, Z. Yan, J. Yao, E. Beitler, Y. Zhu and J. M. Tour, Nature, 2010, 468, 549–552.
27. I. V. Markov, Crystal Growth for Beginners: Fundamentals of Nucleation, Crystal Growth and Epitaxy, World Scientific Publishing Co. Pvt. Ltd., Singapore, 2nd edn, 2003.
28. J. Gao, J. Yip, J. Zhao, B. I. Yakobson and F. Ding, J. Am. Chem. Soc., 2011, 133, 5009–5015.
29. S. Saadi, F. Abild-Pedersen, S. Helveg, J. Sehested, B. Hinnemann, C. C. Appel and J. K. Norskov, J. Phys. Chem. C, 2010, 114, 11221–11227.
30. N. Liu, L. Fu, B. Dai, K. Yan, X. Liu, R. Zhao, Y. Zhang and Z. Liu, Nano Lett., 2011, 11, 297–303.
31. E. Loginova, N. C. Bartelt, P. J. Feibelman and K. F. McCarty, New J. Phys., 2008, 10, 093026.
32. A. Hatton, M. Engstler, P. Leibenguth and F. Mucklich, Adv. Eng. Mater., 2011, 13, 136–144.
33. C. Mattevi, H. Kim and M. Chhowalla, J. Mater. Chem., 2011, 21, 3324–3334.
34. L. F. Mondolfo, B. Vonnegut and H. Chessin, Science, 1972, 176, 695.
35. X. Li, W. Cai, L. Colombo and R. S. Ruoff, Nano Lett., 2009, 9, 4268–4272.
36. C. M. Orofeo, H. Ago, B. Hu and M. Tsuji, Nano Res., 2011, 4, 531–540.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c1nj20695h

---