Discontinuity and misorientation of graphene grown on nickel foil: Effect of the substrate crystallographic orientation

CARBON

9 4 ( 2 0 1 5 ) 1 6 0 –1 7 3

Available at www.sciencedirect.com

ScienceDirect journal homepage: www.elsevier.com/locate/carbon

Discontinuity and misorientation of graphene grown on nickel foil: Effect of the substrate crystallographic orientation Jekaterina Kozlova a, Ahti Niilisk a, Harry Alles a, Va¨ino Sammelselg a b

a,b,*

University of Tartu, Institute of Physics, Ravila 14c, 50411 Tartu, Estonia University of Tartu, Institute of Chemistry, Ravila 14a, 50411 Tartu, Estonia

A R T I C L E I N F O

A B S T R A C T

Article history:

Graphene growth by chemical vapor deposition on low cost metal foils is a promising

Received 2 March 2015

approach for the production of large-scale graphene. However, the precise control of

Accepted 13 June 2015

the uniformity of synthesized mono- and multilayer graphene requires elucidation of the

Available online 19 June 2015

factors affecting deposition and growth. In this study, we investigate the influence of the crystallographic orientation of nickel on multilayer graphene growth using electronbackscatter diffraction, Raman and energy dispersive X-ray spectroscopies, as well as scanning electron and atomic force microscopies. We correlated the discontinuities of the graphene sheets on polycrystalline nickel foils with crystallographic orientations of nickel grains. In addition, we observed indications of misoriented (twisted) multilayer graphene on particular grain orientations. We demonstrate that the Raman signature from these misoriented multilayer graphene areas is highly similar to that previously reported for twisted bilayer graphene. Using microscopy methods, we demonstrated dramatic morphological changes in the nickel substrate induced by graphene growth. Ó 2015 Elsevier Ltd. All rights reserved.

1.

Introduction

Since the isolation of graphene in 2004 [1], the number of proposed applications of graphene is continuing to grow. To take graphene out of the lab for applications on the industrial level, a reproducible method for large-scale synthesis of high-quality homogenous graphene is required. Chemical vapor deposition (CVD) on transition metal surfaces is a potentially useful approach. From an industrial point of view, copper and nickel catalysts are of particular interest due to their availability and low cost. In addition, these catalysts can be readily etched to allow the synthesized graphene to

be transferred to the desired substrate. A self-limiting growth mechanism on copper allows for the production of a predominantly single layer graphene (SLG). Because carbon solubility in copper is negligible, the growth stops when the entire surface is covered with graphene. The synthesis of monolayer graphene on nickel is more challenging due to precipitation-induced growth and is generally achieved on single crystalline substrates in ultra-high vacuum [2,3]. Fewlayer graphene (FLG) and multilayer graphene (MLG) are obtained at standard CVD conditions, where the number of graphene layers can be controlled by varying the thickness of the underlying nickel substrate or the time of hydrocarbon

* Corresponding author at: University of Tartu, Institute of Physics, Ravila 14c, 50411 Tartu, Estonia. E-mail address: [email protected] (V. Sammelselg). http://dx.doi.org/10.1016/j.carbon.2015.06.023 0008-6223/Ó 2015 Elsevier Ltd. All rights reserved.

CARBON

9 4 (2 0 1 5) 1 6 0–17 3

exposure [4,5]. FLG and MLG exhibited good performance as transparent conductive electrodes [6–11], corrosion protection coatings [12,13], and lubricants [14] as well as in sensing applications [15–22] and spintronics [23–25]. For applications where the number of layers is not critical, thicker graphene may offer higher durability and is less affected by the substrate. In addition, MLG has exhibited better performance compared to SLG in some applications, such as gas sensors [21,22] and electrodes for organic photovoltaic cells [6]. Multilayer graphene nanoribbons have a higher electrochemical capacitance compared to single-layer ones [26]. In addition, in contrast to SLG, graphene structures with more layers allow for the creation of hybrid intercalation materials with new appealing properties [27]. Nickel is a widely used metal, and graphene that is directly synthesized on its surface can be used as a protection layer [23,24,28,29]. The strong interaction between graphene and nickel leads to an interesting phenomena. For example, Bodepudi et al. determined that MLG grown on nickel substrates exhibits a large negative magnetoresistive effect, which disappeared after graphene transfer [30]. For most of these applications, it is highly beneficial to avoid discontinuities and obtain uniform graphene coatings on nickel. However, the substrates used for the synthesis are typically polycrystalline, which can have different textures due to manufacturing and subsequent processing. The influence of the substrate crystal structure on graphene growth has previously been demonstrated on copper substrates. Graphene domains tend to exhibit preferred orientations relative to copper substrate [31,32], and the shape of graphene nuclei reflects the symmetry of the underlying crystallographic plane [32,33]. Despite the fact that individual graphene islands can be grown continuously across Cu grain boundaries [34], the change in the graphene domain shapes across different Cu grains on foils with different textures has been demonstrated for SLG under low-pressure chemical vapor deposition conditions [31] and for FLG domains under atmospheric pressure during chemical vapor deposition [35]. However, additional studies reported that graphene domains can grow into different morphologies on the same grain [36,37]. Wood et al. and Zhao et al. reported different graphene growth rates, which were dependent on the copper orientation, and the fastest growth rate was achieved on the (1 1 1) surface [38,39]. In addition, graphene coverage is influenced by the orientation of the copper substrate. Growth on the (1 1 1) face was complete but graphene on some other orientations, such as (3 1 0), was in the form of discrete islands [38,39]. In addition, the quality of the obtained graphene tended to correlate with the orientation of the substrate where graphene on the Cu(1 1 1) plane had a higher monolayer percentage. In addition, the bilayer and few-layer graphene was produced on the (1 0 0) [31,40] and (1 1 0) planes [39] as well as on high-index surfaces [39,40]. The results on the different faces of copper foil are consistent with the results reported by Ogawa et al. who compared graphene growth on heteroepitaxial Cu(1 1 1) and Cu(1 0 0) films on MgO single crystal substrates and determined that the growth on Cu(1 1 1) oriented films led to better graphene quality [41].

161

The comparison of epitaxial Cu(1 1 1) film on the c-plane sapphire to crystallographically more diverse copper film on Si/SiO2 [42] and Cu foil [43] revealed a similar trend. In comparison to copper, there are limited reports discussing the influence of nickel foil orientation on graphene growth [44,45]. In this study, by employing electronbackscatter diffraction (EBSD), scanning electron microscopy (SEM), energy-dispersive X-ray microanalysis (EDX), atomic force microscopy (AFM) and Raman scattering techniques, we demonstrate the dependence of the crystallographic orientation of nickel grains on the continuity of graphene growth. In addition, the domain structures of the packing of the graphene sub-layers in the multi-layer graphene stacks that distinguish the MLG from the graphite were identified. Besides, considerable restructuring of the Ni foil surface was observed due to the graphene growth.

2.

Materials and methods

2.1.

Graphene synthesis

Graphene synthesis was carried out on 25 lm thick Ni foils (99.9%, Strem Chemicals, Inc.) using a low-pressure CVD method in a hot-wall laboratory reactor from methane as a carbon source. The CVD reactor was comprised of a quartz tube with length of 500 mm and diameter of 25 mm, placed inside resistively heated furnace having length of 300 mm. The temperature of a quartz sample holder, placed in the middle of the quartz tube, was recorded using a thermocouple. The reactor was equipped with three digital mass-flow controllers (Smart-Trak 2, Sierra Instruments) for gas lines of Ar, 10% CH4/Ar and H2. Water cooling of the reactor flanges was provided via closed system. Prior to the growth, the nickel foils were annealed at 1000 °C for 1 h in a flow of H2 and Ar gases (both 99.999, AGA Estonia) with flow rates of 125 and 100 sccm, respectively. Then, 140 sccm of 10% CH4 in Ar gas (AGA Estonia) was introduced into the reactor for different growth periods. The total pressure in the CVD chamber was maintained at 3 mbar. Finally, the sample was rapidly cooled by moving the furnace away from the sample position. The average cooling rate achieved this way was of 60 °C/min, where the temperature drop up to 700 °C was particularly fast of 300 °C/min. These cooling conditions are denoted in the text as a ‘‘fast cooling’’ process. The CH4 and Ar/H2 flows were maintained during the cooling to 200 °C. The same above-mentioned cooling conditions were used for all the samples unless otherwise stated. For comparison, we also performed a slow-cooling process with an average cooling rate of 7 °C/min. First, the sample was cooled down to approximately 700 °C at a rate of 4 °C/min, afterward, the furnace was switched off, and the sample was left to cool to room temperature. These cooling conditions are denoted in the text as a ‘‘slow cooling’’ process. Some samples with deposited graphene were treated in low pressure Ar plasma using plasma cleaner Femto-1 (Diener Electronic GmbH) for removing partly or fully the graphene coating.

162

2.2.

CARBON

9 4 ( 2 0 1 5 ) 1 6 0 –1 7 3

Sample characterization

The samples were characterized with a Raman spectrometer (inVia, Renishaw) using a laser excitation wavelength of 514.5 nm, a high-resolution scanning electron microscope with a Schottky-type electron emitter (HR-SEM, Helios NanoLab 600, FEI Company), and an energy-dispersive X-ray spectrometer (EDX, INCA Energy 350, Oxford Instruments) connected to the same microscope. HR-SEM micrographs were recorded using an acceleration voltage of 2 kV. The crystallographic orientation of the grains of the nickel substrates was determined by electron backscatter diffraction (EBSD) measurements using a different scanning electron microscope (EVO MA 15, Zeiss) equipped with an EBSD detector (HKL Nordlys, Oxford Instruments). The EBSD data were post-processed using the Aztec (Oxford Instruments) and CHANNEL 5 (HKL) software programs. For the EBSD measurements, the accelerating voltage of the primary beam was 20 kV. To enhance the detection sensitivity of the EBSD patterns, the samples were inclined approximately 70° relative to the electron beam. The surface morphology of the nickel was studied using AFM (AutoProbe CP II, Veeco) by simultaneously recording the topography and error-signal modes. The images presented in this paper are recorded in the error signal mode due to its higher sensitivity to edges, which allowed for better visualization of small surface features on significantly rough surfaces. The dimensions of the surface features were measured on the corresponding topography images.

3.

Results and discussion

3.1. Pre-annealing influence on the microstructure of Ni-foil The microstructure of the nickel foil before and after annealing prior to exposure to methane has been explored by SEM and EBSD (Fig. 1). As shown in the SEM micrograph in Fig. 1a, the as-received nickel foil has a structure with a grain size of approximately 5 lm. The striations resulting from the rolling process of the foil are visible on the surface. The EBSD color map in Fig. 1b shows the crystallographic orientation of the outermost surfaces of the nickel grains/crystallites relative to the sample surface normal direction (ND). The colors in the corners of the color-triangle (Fig. 1e) show the lowindex surfaces, and the colors between them indicate stepped low-cut or vicinal surfaces. The black or white points indicate that it was not possible to determine the crystallographic orientation of the crystallites surfaces in these points. As shown in Fig. 1b, the nickel surface has a wide range of crystallographic orientations. However, grains with an orientation close to the Ni(1 1 1) orientation are more common. The nickel crystal has a face-centered cubic unit cell, and the Ni(1 1 1) surface, which possesses the lowest surface energy, also has the smallest lattice mismatch with graphene and is considered to be the most suitable for graphene growth. After 1 h of annealing in an Ar/H2 flow, the SEM micrograph of the foil indicates a much smoother surface morphology (Fig. 1c). The deepest rolling striations are still visible but overall, they are

Fig. 1 – SEM images of nickel foil (a) before and (c) after annealing. EBSD orientation map of nickel foil (b) before and (d) after annealing as well as the (e) color key associated with different crystalline orientations.

CARBON

9 4 (2 0 1 5) 1 6 0–17 3

less pronounced. The nickel foil is largely recrystallized, and the size of the crystallites has significantly increased. As shown in Fig. 1d, annealing did not have a significant effect on the crystallographic orientation of the grains. In addition to the Ni(1 1 1) face, facets of higher surface energy remain on the nickel surface after annealing.

3.2.

Investigation of the initial growth stage of graphene

To investigate how graphene growth initiates and progresses, SEM images were recorded after different methane exposure times at 1000 °C (Fig. 2). The high graphene growth temperature was to ensure a low number of graphene defects [46]. The graphene growth starts as carbon atoms combine with the formation of carbon clusters and once the cluster size exceeds the critical size the nucleation starts. As it has been shown in the theoretical work of Gao et al. [47] the graphene growth on nickel starts from the carbon monomers or dimers aggregation first to linear chains and further, by increased number of carbon atoms in the initial cluster to sp2 rings, which serve as seeds for graphene growth. However, graphene nucleation on terrace can take place, and even easier at higher temperatures, the nucleation near the step edges, kinks and other lattice defects is energetically more favorable. The exceptionally low secondary electron yield of graphene [48] allows distinguishing even a monolayer graphene island from the metal surface. Fig. 2a and b shows SEM images of the nickel surface after 45 s of methane exposure at 1000 °C. Two types of islands were observed as follows: larger ones with average dimensions of several micrometers with some being larger than 10 lm and numerous small islands with dimensions less than several hundreds of nm. Small graphene islands can be divided into two groups. In some substrate grains they exhibit somewhat elongated shape with the prolongation direction parallel to the grain longest side. At the same time in other parts of the substrate nearly circular shape islands irrespective of the orientation of the underlying nickel grain can be seen. Also, the larger islands of both regions can be interpreted as circular. Thus, at high temperatures of the graphene growth carbon surface diffusion rate can be interpreted as isotropic, at least at the island edges far enough from the boundaries of the substrate grains. For the islands nucleated near or on the grain boundaries, elongation of graphene islands was observed along the boundary. However, the nucleation of graphene islands is not completely random. Although some of the grains have a large number of nucleation centers, other grains can be completely uncovered. The nonuniform contrast within the graphene islands indicates the variation in the thickness. Therefore, at the beginning of graphene growth, the islands consist of more than one layer despite the fact that most of the surface is still uncovered. Darker areas, which correspond to thicker graphene, were observed in the middle of the islands and more frequently observed on the island periphery. When the exposure time to methane was increased to 2 min at 1000 °C, approximately half of the surface was covered (Fig. 2c and d). The size of the islands increased such that their coalescence was observed in numerous areas. In addition, there are a large number of small islands that may act as secondary nucleation centers. With the increased

163

coverage, the formation of meandering steps was observed on the surface of the nickel foil. The nickel surface morphology was studied in more detail by AFM (Fig. 8) and will be discussed below. After 5 min of exposure, most of the nickel surface is covered with graphene (Fig. 2e and f). However, some uncovered areas, which are indicated by the brightest regions, still remain on the surface. The borders of the uncovered areas tend to coincide with the nickel grain boundaries. These results confirm that the crystallographic orientation of the nickel grains has an influence on graphene growth. It is interesting to note that the areas covered with graphene exhibit a wrinkled morphology, which was not observed in the initial stages of graphene growth. The stepped pattern of the uncovered areas also became more pronounced. The graphene appears to spill over to uncovered grains from wellcovered grains along these steps. Even after 1 h of exposure to methane at 1000 °C, discontinuities in the coating remain (Fig. 2g and h). Because the topography and electron channeling contrast of the underlying nickel substrate can affect how dark or bright certain areas appear in the SEM images, both EBDS and EDX mappings were conducted to investigate the carbon distribution on the nickel surface (Fig. 3). To enhance the X-ray microanalysis surface sensitivity, low acceleration voltage of 2 kV was employed for the EDS analysis. Quantitative maps of C Ka and Ni La are shown in Fig. 3c and d, respectively. The bar provided in the lower portion of the images represents the percentage of the element within the electron-excited near-surface volume. The EDX mapping confirmed that the carbon distribution on the surface is non-uniform, which is consistent with the SEM images. In addition, from the EDX maps, an increased carbon signal was observed between nickel grains due to carbon precipitation or its faster diffusion at grain boundaries.

3.3.

Study of the graphene coverage of the Ni-foil

To study the dependence of graphene coverage on nickel grain orientation, EBSD mapping was performed after longer periods of CVD growth. Figs. 4 and 5 show a comparison of the SEM images and EBSD maps recorded of the same areas after methane exposure at 1000 °C for 10 min and 1 h, respectively. After 5 min (Figs. 2e, f, and 3) and 10 min (Fig. 4), a noticeable difference in the graphene coverage on different grains was observed. The grains with a crystallographic orientation close to Ni(1 1 1) are well covered. However, noncontinuous graphene coverage was observed on grains oriented close to Ni(0 0 1) and especially on higher index surfaces (e.g., rose colored grains with an orientation roughly in the middle between Ni(1 1 1) and Ni(0 0 1) or yellow colored grains with a surface orientation that was roughly in the middle between Ni(1 1 1) and Ni(1 0 1)). Graphene appears to extend to these faces from neighboring grains, forming long dendritic branches oriented along the steps on the nickel substrate (see in Figs. 2d, 3a, 4c, and S4c, the latter in the Supplementary materials). In contrast, continuous graphene coverage was obtained on light blue colored high-index grains with an orientation in the middle between Ni(1 1 1) and Ni(1 0 1). The dependence of graphene growth on the nickel crystalline orientation has been reported by Takahashi et al.

164

CARBON

9 4 ( 2 0 1 5 ) 1 6 0 –1 7 3

Fig. 2 – SEM images of the surface coverage with graphene film after the CVD process with (a and b) 45 s, (c and d) 2 min, (e and f) 5 min, and (g and h) 1 h of methane exposure, all at 1000 °C and fast cooling.

CARBON

9 4 (2 0 1 5) 1 6 0–17 3

165

Fig. 3 – Comparison of the SEM image (a) and EBSD map (b) as well as the EDX quantmap of carbon (c) and nickel (d) from the same location on the sample after the CVD process with 5 min of methane exposure at 1000 °C and fast cooling. The bars on the EDX quantmaps indicate the concentration of the analyzed element in the analyzed volume.

[44], who studied in situ graphene formation by annealing nickel foils pre-covered with disordered graphene layers in a SEM chamber. The observed difference in graphene formation on the (0 0 1)-oriented grains was due to a larger binding energy of the carbon atoms to this surface compared to the (1 1 1) surface. However, the situation may be more complicated because during graphene growth, obvious surface restructuring occurs in some areas including the generation of steps that introduce additional surface orientations. In addition, it may be possible that some metal surface areas suffer from reconstruction. Therefore, the atomic structure of the metal surface may differ from that of the bulk. However, the latter effect, which could influence graphene nucleation and growth, cannot be studied with the EBSD method. As shown in Figs. 3b, 4b, and 5d, the grains with surface orientations near to the (1 1 1) exhibit slightly different shades of blue with an average misorientation angle of approximately 10°. Interestingly, when the orientation of the grain deviates less from the ideal (1 1 1) orientation, a thinner graphene coating is formed on the grain (see Fig. S1 in the Supplementary materials). After the increase in the methane exposure time to 1 h at 1000 °C, the grains with a (0 0 1) orientation as well as the high-index surface areas are nearly covered with graphene. However, some small holes or discontinuities in the graphene coating were still observed, especially if the area of these grains was quite large (Fig. 5a and b). With this extended exposure time, the peculiar coverage of the grains with a sharp Ni(1 1 1) orientation were more obvious. These grains exhibit a distinctive

graphene coverage pattern and are characterized by periodic bands of thinner and thicker graphene. Next, to improve the coating homogeneity, a slow cooling process was utilized. As shown in Fig. 6a, the slow cooling yielded the opposite effect. The SEM image shows areas of apparently very thick graphitic coating with a wrinkled morphology and areas with significantly thinner graphene coverage. The EBSD map (Fig. 6b) indicates that the areas with a thinner coverage have a (1 1 1) orientation. Therefore, surface restructuring during graphene growth may cause carbon transfer from the (1 1 1)-oriented grains to the neighboring grains (e.g., by enhanced surface diffusion). The overall effect of the slow cooling rate is a significant increase in the thickness inhomogeneity of graphene coating. The grains with orientation close to the Ni(1 1 1) have thinner graphene coating and the thinnest coating is observed on grains with ideal Ni(1 1 1) orientations. Grains with other orientations are covered with relatively thick multilayer graphene.

3.4.

Raman study of multi-layer graphene

In Fig. 7, we show typical Raman spectra of the graphene collected on the nickel foil after 30 min of methane exposure at 1000 °C and fast cooling where a nearly continuous coverage was obtained. As shown in the overview spectrum in Fig. 7a, the intensity value of the G band (1583 cm1) is higher than that of the 2D band (2715 cm1), and the last band has an asymmetric shape. This band can be divided into two

166

CARBON

9 4 ( 2 0 1 5 ) 1 6 0 –1 7 3

Fig. 4 – Comparison of SEM images (a and c) and EBSD orientation maps (b and d) taken from the same area in two locations showing the dependence of coverage on grain orientation after 10 min of methane exposure at 1000 °C and fast cooling. Lorentzian curves with maxima at 2734 and 2703 cm1 (not shown), and a band splitting of D2D = 31 cm1. Based on the relationship of D2D to the number of graphene layers [49], in the AB-stack approximation, the number of Raman-active graphene layers was estimated to be ~10. This AB-stack approximation was confirmed by the existence of weak spectral bands between 1650 and 1800 cm1 (right inset, Fig. 7a), which correspond to the out-of-plane combination modes of the longitudinal optical (LO), out-of-plane acoustic (ZA), outof-plane optical (ZO 0 ) and an overtone of another out-ofplane breathing vibration mode (ZO): (LO + ZA, LO + ZO 0, 2ZO). These modes are specific to the AB-stacked graphene [50–52]. In addition, the Raman features in the range of 80– 130 cm1 corresponded to the ZO 0 breathing modes, and facilitated by rotational disorder [53,54], are missing (left inset, Fig. 7a). All of these Raman features indicate the presence of multilayer graphene that primarily consists of graphitelike AB stacked layers. In addition, in certain regions over the sample, the spectra were similar to that shown in Fig. 7b. In this case, the Raman spectrum consists of a very intense and narrow 2D band (2707 cm1) as well as a low-intensity G band (1582 cm1). The integral intensities ratio (A2D/AG) can reach a value as high as 15 or greater (e.g., in Fig. 7b, A2D/AG = 6.7). Although the spectrum in Fig. 7b is similar to the spectrum of monolayer graphene, no decrease in the carbon concentration was observed in these areas in the EDX maps compared to the neighboring grains (Supplementary materials, Fig. S2).

The rotation of the adjacent graphene planes can result in the Raman spectra of thicker graphene being similar to that obtained for a single layer graphene. This behavior was primarily reported for misoriented (non-AB stack) bilayer systems [55–57]. However, this peculiarity was also predicted for misoriented multilayer graphene [58] and demonstrated on multilayer graphene grown on the C-face of SiC [59,60]. In contrast to the AB stack (Fig. 7a) in the spectra obtained in these distinctive areas, the bands connected with the out-of-plane modes in the 1650–1800 cm1 spectral range are missing, and the bands connected to the in-plane combination modes of the longitudinal optical (LO), transverse acoustic (TA), transverse optical (TO) and longitudinal acoustic (LA) vibration modes (LO + TA, TO + LA, LO + LA, TO + TA) between 1800 and 2250 cm1 are well pronounced (right inset, Fig. 7b). Along with the high value of the A2D/AG ratio, this effect is unique to weakly bounded layers inside the graphene system with stacking disorder between the layers [50,51]. Additional information on the stacking order can be obtained by the bands connected to the rotation modes (R) [53,61,62], which appear in the lower energy region of the spectrum between defect band D at 1357 cm1 and band G at 1582 cm1 (right inset, Fig. 7b). In this spectral range, at least two bands assigned to R modes were isolated at 1383 and 1483 cm1. The corresponding rotation angles (h) were estimated to be approximately 26° and 13°, respectively, based on the xR = xR(h) relationship. In the spectral range of 80–130 cm1 belonging to the ZO 0 breathing modes [53,54]

CARBON

9 4 (2 0 1 5) 1 6 0–17 3

167

Fig. 5 – Comparison of SEM images (a and c) and EBSD orientation maps (b and d) taken from the same area in two locations showing the dependence of coverage on the grain orientation after 1 h of methane exposure at 1000 °C and fast cooling.

Fig. 6 – Comparison of the SEM image (a) and EBSD orientation maps (b) from the same location after the CVD growth with a slow cooling process. The white color on the EBSD map indicates areas where the EBSD signal was not obtained. (A color version of this figure can be viewed online.)

several overlapping Lorentzian peaks were observed. The most intense peaks of them are located at 106, 115 and 121 cm1, which indicate the presence of several rotated graphene layers that have slightly different breathing frequencies. Based on the recent estimates [63,64], these peaks most likely correspond to a combination of vibrational modes, which involve both the ZO 0 breathing and shear (C)

modes, (ZO 0 + C) in the misoriented portion of the multilayer graphene. The most spread on the sample were the areas with mixed structure, which consist of AB-stacked layers along with rotationally faulted graphene layers. A typical set of Raman scattering spectra for this type of domain is shown in Fig. 7c. The Raman spectrum displays two intense commonly

168

CARBON

9 4 ( 2 0 1 5 ) 1 6 0 –1 7 3

Fig. 7 – Raman spectra from different points of a graphene sample prepared with 30 min of methane exposure at 1000 °C and fast cooling: (a) MLG with AB stacking, (b) misoriented MLG, (c) MLG with a mixed structure of AB and rotationally faulted graphene layers stacking, and (d) EBSD orientation map showing the orientation dependence (color code given in Fig. 1e) of the measured Raman signal – symbol D refers to the spectrum given in (a),  refers to the spectrum given in (b), s refers to the spectrum given in (c), and the Raman maps of the peak ratio A2D/AG (presented in their own color scales) are taken with a step of 10 lm from the areas marked with dashed rectangles on the EBSD map.

symmetrical bands G and 2D located at 1582 and 2707 cm1, respectively. The intensity of the G band was less than the intensity of the 2D band. The A2D/AG integral intensity ratio was 5, which is quite similar to that for the Raman spectrum of exfoliated SLG. However, the widths of the bands were too large for single layer graphene, and the broad 2D band can be divided into two sub-bands at 2695 and 2716 cm1 with an interval between the maxima (2D band splitting) of 21 cm1,

giving in the AB-stack approximation for the number of graphene layers on Ni substrate approximately 3 [49]. Indeed, the AB stack was observed after the LO + ZO 0 peak at 1752 cm1 (right upper inset, Fig. 7c). However, the use of the AB stack approximation to estimate the number of layers is inadequate because additional details in the spectrum exist, which demonstrates the substantial role of the rotationally faulted graphene layers. In the same inset, other

CARBON

9 4 (2 0 1 5) 1 6 0–17 3

169

detected including a broadened structure below 800 cm1 and peaks at 820 and 841 cm1 (lower right inset, Fig. 7c). Two former peculiarities agree well with the results that in the spectrum, two R bands at 1421 and 1479 cm1 exist and affirm the appearance of two rotation angles at 20° and 14° for multilayer graphene film. The latter peak at 841 cm1 confirms the additional twisted graphene layer(s) with an angle of approximately 10°. The corresponding rotation mode was not observed in the Raman spectrum even though it should appear in the region above 1500 cm1 where it coincides with the intense G band. Fig. 7d shows the correlation between crystal orientation mapping and the obtained Raman spectra. The Raman signal with the unusually high 2D/G ratio was primarily observed on the grains with an orientation roughly in the middle between Ni(0 0 1) and Ni(1 1 1) (see the color key in Fig. 1). The triangular, rhombic and circular symbols superimposed on the EBSD orientation map correspond to the spots locations where spectra similar to Fig. 7a–c, respectively, were measured. Therefore, the Raman characterization of CVD graphene on the Ni catalyst indicates that there is a great variety of multilayer graphene domains with very different stacking orders that distinguish the MLG from the graphite. The exact number of graphene layers cannot be precisely determined due to the possible existence of non-Raman active layers [44] and the presence of rotational disorder between the graphene layers. The results from the Raman scattering spectra of CVD grown graphene on Ni foil unambiguously indicate the existence of AB stacked and rotated layers that are located together as well as separately inside the multilayer graphene domains. The rotation of the graphene layers affects the electronic properties of the synthesized graphene sheets. As shown for the large rotation angles, the misoriented graphene system can exhibit apparent SLG behavior [67]. In particular, the rotation disorder induced electronic decoupling between layers can lead to high charge carrier mobility even in multilayer structures [68]. Fig. 7 (continued)

3.5. AFM and SEM investigations of the Ni foil surface morphology

details in Raman spectrum are shown. The in-plane vibration modes (i.e., LO + TA, TO + LA, LO + LA, TO + TA [50,51,65]) and two sharp lower energy peaks due to the rotation modes at 1421 and 1479 cm1 were observed. Based on the given frequency values for the rotation modes, the corresponding twisting angles between the stacking graphene layers were determined to be approximately 20° and 14°, respectively. The role of rotational stacking was further approved by Raman analysis of the breathing modes ZO 0 and ZO (left and lower right insets, Fig. 7c). As in the previous case (Fig. 7b), in the spectral region of the ZO 0 breathing modes, a structure exists, which can be divided into three Lorentzian-shape bands (at 107, 115 and 122 cm1) of several rotated graphene layers in MLG. The ZO breathing mode is characterized by the Raman peak at approximately 800 cm1 [66], and its exact position has a tight relationship to the spectral position of the rotation mode R and therefore to the rotation angle h [54]. On the high background, due to the Ni substrate, some very weak peculiarities in this region were

As previously mentioned, morphology changes in the nickel surface were observed during graphene growth. In Fig. 8a, the AFM image of the Ni foil after only the annealing step, performed at the same conditions as prior the all graphene syntheses, and during the samples cooling the gases flow was the same as used during the annealing, displays a smooth surface with flat terraces separated by steps of subnanometer height. A brief exposure to methane for 2 min led to the appearance of periodic surface modulations that are visible in the areas not covered by graphene (Fig. 8b). These wavy surface modulations appear well ordered with a peak height of 61 to 10 nm as shown in Fig. 8c and a period from 30 to more than 100 nm depending on the particular place. However, areas between the graphene islands retained their smooth morphology. Apparently, graphene islands impede the flow of steps on the catalyst surface, which could be induced by surface energy anisotropy. In addition, chemisorbed molecules can facilitate surface diffusion by weakening the bonds between neighboring substrate atoms [69,70]. As the step flow

170

CARBON

9 4 ( 2 0 1 5 ) 1 6 0 –1 7 3

Fig. 8 – AFM images of the surface of the nickel foil: (a) after annealing, (b) after CVD growth (2 min of methane exposure at 1000 °C, fast cooling) between the graphene islands, (c) after CVD growth (2 min of methane exposure at 1000 °C, fast cooling), and (d) after the graphene removal with plasma cleaning. The polynomial background was subtracted in all of the images. (A color version of this figure can be viewed online.)

Fig. 9 – SEM images of the surface of the Ni foil with graphene partially removed by Ar plasma cleaning: (a) CVD growth with 5 min exposure to methane at 1000 °C and slow cooling process with methane flux and (b) CVD growth with 1 h of exposure to methane at 1000 °C and fast cooling.

proceeds, the velocity of the step is affected by the graphene nucleation events on the terraces in front of the steps, which leads to different step velocities and bunching of the steps. The graphene covered areas also exhibit a rougher morphology (Fig. 8c and d). In contrast to the bare nickel surface between the graphene islands, these areas lacked the ordered pattern or initial flat relief and exhibited an irregular bumpy surface. To determine if the observed corrugated morphology was caused by changes in the nickel surface or graphene itself, i.e., by the difference in the heat expansion coefficient between graphene and nickel the graphene was removed

from the surface using argon plasma cleaning. As shown in Fig. 8d, after graphene removal, the surface exhibits the same roughened morphology, which indicates that the graphene growth induces structural transformations in the nickel surface. To determine the nickel substrate surface change dependence on the graphene growth period two samples were studied with SEM after partly removed graphene coating using the plasma treatment. In Fig. 9a is shown in the first sample with thick graphene coverage, which was obtained after 5 min of growth followed by slow cooling process, exhibits pronounced

CARBON

9 4 (2 0 1 5) 1 6 0–17 3

surface changes characterized by the high steps with rough edges. The second sample was prepared with prolonged graphene growth (1 h), and the graphene layer was partially removed from it by 20 min of plasma treatment (Fig. 9b). The SEM image reveals a surface covered in steps meandering by graphene islands. In some places, the moving steps were unable to overcome precipitates bent around the obstacle on both sides forming nearly circular closed step loops. Thus, wave-like structures (meandering steps) can be a result from the disturbances created by nucleation events on initially straight steps. In graphene covered areas the surface diffusion as well as sublimation and redeposition of substrate atoms is inhibited, which hamper the advancement of the steps leading to decreased velocity compared to the steps on open surface. Phase coherence of the steps develops later due to the effective step-step repulsion. A driving factor responsible for morphological instability is asymmetric step kinetics induced by carbon adsorption and nucleation of graphene islands as in annealed only substrates we did not find areas with such kind of surface morphology. The step bunches seem to favor graphene nucleation and subsequent expanding graphene to the whole terrace. Thus, the surface changes increase as the graphene growth time increases. During graphene growth for a long duration of time at high temperatures, the reorientation of the metal grains accompanied by restructuring of their surface may contribute to the changes in the surface morphology. In addition, based on the strong bright contrast along the edges between graphene free areas and areas where graphene was not completely removed by plasma cleaning, a height difference was observed between these areas. The areas where the graphene was not completely removed were most likely covered with thicker graphene, and they appear to be lower compared to the areas, which had fewer graphene layers. This observation strengthens the previous suggestion that the nickel surface is mobile during graphene growth. The constantly changing surface during graphene growth may result in incomplete coverage of the metal surface with graphene even during a prolonged growth process and the formation of rotational disorder in the synthesized graphene coatings.

4.

Conclusions

In this study, we demonstrated that the growth of graphene layers on commercial polycrystalline Ni foils using lowpressure CVD method depends strongly on the surface orientation of the substrate monocrystalline grains. The synthesized graphene was examined at different growth stages. The graphene growth dependence on the orientation leads to discontinuities in the graphene coverage even after prolonged growth times when a considerable portion of the substrate was covered with multilayer graphene. The orientations close to Ni(1 1 1) are preferential for graphene growth and exhibited better graphene coverage at a given time compared to orientations close to Ni(0 0 1) and higher index surfaces. In addition, the growth on nickel grains with sharp (1 1 1) orientation results in a graphene coating with fewer layers, which was especially noticeable for the growth process at slower cooling rates.

171

After a detail investigation of the Raman spectral features, we identified areas with misoriented multilayer graphene that clearly distinguish the CVD grown multilayer graphene from a graphite sheet. In addition, the relationship between the misorientation of neighboring layers in multilayer graphene and the surface orientation of the nickel substrate was demonstrated, especially for the orientation roughly in the middle between (1 1 1) and (1 0 0) of nickel, which enhances the misorientation (twisting) of the graphene layers in the multilayer graphene stack. The latter can lead to multilayer graphene with electronic properties similar to singlelayer graphene because the interaction between the layers is smaller than that in the ordered stack. Finally, remarkable morphological changes in the nickel surface were observed during graphene growth, which have been mostly overlooked and should be considered for the design of graphene synthesis on nickel.

Acknowledgments Prof. Kalle Kirsima¨e from the Institute of Ecology and Earth Sciences, University of Tartu, is gratefully acknowledged for allowing access to the Zeiss EVO MA 15 and for aiding in the EBSD measurements. In addition, we wish to thank Mr. Lauri Aarik for construction of the CVD reactor used for graphene growth. This study was partially supported by the European Union through the European Social Fund (Grant MTT1) and FP7 programme under Grant Agreement No. 604391 Graphene Flagship and by the Estonian Ministry of Education and Research (Projects IUT2-24 and TK117).

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.carbon. 2015.06.023.

R E F E R E N C E S

[1] Novoselov KS, Geim AK, Morozov SV, Jiang D, Zhang Y, Dubonos SV, et al. Electric field effect in atomically thin carbon films. Science 2004;306:666–9. [2] Addou R, Dahal A, Sutter P, Batzill M. Monolayer graphene growth on Ni (1 1 1) by low temperature chemical vapor deposition. Appl Phys Lett 2012;100. 021601-1–3. [3] Iwasaki T, Park HJ, Konuma M, Lee DS, Smet JH, Starke U. Long-range ordered single-crystal graphene on high-quality heteroepitaxial Ni thin films grown on MgO (1 1 1). Nano Lett 2010;11:79–84. [4] Weatherup RS, Dlubak B, Hofmann S. Kinetic control of catalytic CVD for high-quality graphene at low temperatures. ACS Nano 2012;6:9996–10003. [5] Reina A, Thiele S, Jia X, Bhaviripudi S, Dresselhaus MS, Schaefer JA, et al. Growth of large-area single-and bi-layer graphene by controlled carbon precipitation on polycrystalline Ni surfaces. Nano Res 2009;2:509–16. [6] Choe M, Lee BH, Jo G, Park J, Park W, Lee S, et al. Efficient bulk-heterojunction photovoltaic cells with transparent multi-layer graphene electrodes. Org Electron 2010;11:1864–9.

172

CARBON

9 4 ( 2 0 1 5 ) 1 6 0 –1 7 3

[7] Lee S, Jo G, Kang SJ, Wang G, Choe M, Park W, et al. Enhanced charge injection in pentacene field-effect transistors with graphene electrodes. Adv Mater 2011;23:100–5. [8] Kim BJ, Mastro MA, Hite J, Eddy CR, Kim J, Kim J. Transparent conductive graphene electrode in GaN-based ultra-violet light emitting diodes. Opt Express 2010;18:23030–4. [9] Jo G, Choe M, Cho CY, Kim JH, Park W, Lee S, et al. Large-scale patterned multi-layer graphene films as transparent conducting electrodes for GaN light-emitting diodes. Nanotechnology 2010;21. 175201-1–6. [10] Wang G, Kim Y, Choe M, Kim TW, Lee T. A new approach for molecular electronic junctions with a multilayer graphene electrode. Adv Mater 2011;23:755–60. [11] Choi YY, Kang SJ, Kim HK, Choi WM, Na SI. Multilayer graphene films as transparent electrodes for organic photovoltaic devices. Sol Energy Mater Sol Cells 2012;96:281–5. [12] Prasai D, Tuberquia JC, Harl RR, Jennings GK, Bolotin KI. Graphene: corrosion-inhibiting coating. ACS Nano 2012;6:1102–8. [13] Nayak PK, Hsu CJ, Wang SC, Sung JC, Huang JL. Graphene coated Ni films: a protective coating. Thin Solid Films 2013;529:312–6. [14] Kim KS, Lee HJ, Lee C, Lee SK, Jang H, Ahn JH, et al. Chemical vapor deposition-grown graphene: the thinnest solid lubricant. ACS Nano 2011;5:5107–14. [15] Ma J, Xuan H, Ho HL, Jin W, Yang Y, Fan S. Fiber-optic Fabry– Pe´rot acoustic sensor with multilayer graphene diaphragm. IEEE Photon Technol Lett 2013;25:932–5. [16] Zhu SE, Ghatkesar MK, Zhang C, Janssen GCAM. Graphene based piezoresistive pressure sensor. Appl Phys Lett 2013;102. 161904-1–3. [17] Goh MS, Pumera M. Graphene-based electrochemical sensor for detection of 2,4,6-trinitrotoluene (TNT) in seawater: the comparison of single-, few-, and multilayer graphene nanoribbons and graphite microparticles. Anal Bioanal Chem 2011;399:127–31. [18] Jiang S, Wang X, Shi S, Yuan J, Fang J, Gao C, et al. Simulation of deflection and stress for MEMS pressure sensor based on graphene. 14th International Conference on Electronic Packaging Technology. Dalian, China: IEEE; 2013. p. 462–6. [19] Huang Y, Dong X, Liu Y, Li LJ, Chen P. Graphene-based biosensors for detection of bacteria and their metabolic activities. J Mater Chem 2011;21:12358–62. [20] Goh MS, Pumera M. The electrochemical response of graphene sheets is independent of the number of layers from a single graphene sheet to multilayer stacked graphene platelets. Chem Asian J 2010;5:2355–7. [21] Antonova IV, Mutilin SV, Seleznev VA, Soots RA, Volodin VA, Prinz VY. Extremely high response of electrostatically exfoliated few layer graphene to ammonia adsorption. Nanotechnology 2011;22. 285502-1–5. [22] Choi H, Jeong HY, Lee DS, Choil CG, Choi SY. Flexible NO2 gas sensor using multilayer graphene films by chemical vapor deposition. Carbon 2013;14:186–9. [23] Dlubak B, Martin MB, Weatherup RS, Yang H, Deranlot C, Blume R, et al. Graphene-passivated nickel as an oxidation-resistant electrode for spintronics. ACS Nano 2012;6:10930–4. [24] Shiraishi M, Ohishi M, Nouchi R, Mitoma N, Nozaki T, Shinjo T, et al. Robustness of spin polarization in graphene-based spin valves. Adv Funct Mater 2009;19:3711–6. [25] Goto H, Kanda A, Sato T, Tanaka S, Ootuka Y, Odaka S, et al. Gate control of spin transport in multilayer graphene. Appl Phys Lett 2008;92. 212110-1–3. [26] Goh MS, Pumera M. Multilayer graphene nanoribbons exhibit larger capacitance than their few-layer and single-layer

[27]

[28]

[29]

[30]

[31]

[32] [33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

graphene counterparts. Electrochem Commun 2010;12:1375–7. Zhan D, Sun L, Ni ZH, Liu L, Fan XF, Wang Y, et al. FeCl3-based few-layer graphene intercalation compounds: single linear dispersion electronic band structure and strong charge transfer doping. Adv Funct Mater 2010;20:3504–9. Dedkov YS, Fonin M, Laubschat C. A possible source of spinpolarized electrons: the inert graphene/Ni (1 1 1) system. Appl Phys Lett 2008;92. 052506-1–3. Krishnamurthy A, Gadhamshetty V, Mukherjee R, Chen Z, Ren W, Cheng HM, et al. Passivation of microbial corrosion using a graphene coating. Carbon 2013;56:45–9. Bodepudi SC, Singh AP, Pramanik S. Current-perpendicularto-plane magnetoresistance in chemical vapor depositiongrown multilayer graphene. Electronics 2013;2:315–31. Murdock AT, Koos A, Britton TB, Houben L, Batten T, Zhang T, et al. Controlling the orientation, edge geometry, and thickness of chemical vapor deposition graphene. ACS Nano 2013;7:1351–9. Jacobberger RM, Arnold MS. Graphene growth dynamics on epitaxial copper thin films. Chem Mater 2013;25:871–7. Meca E, Lowengrub J, Kim H, Mattevi C, Shenoy VB. Epitaxial graphene growth and shape dynamics on copper: phase-field modeling and experiments. Nano Lett 2013;13:5692–7. Yu Q, Jauregui LA, Wu W, Colby R, Tian J, Su Z, et al. Control and characterization of individual grains and grain boundaries in graphene grown by chemical vapour deposition. Nat Mat 2011;10:443–9. Wu YA, Robertson AW, Schaffel F, Speller SC, Warner JH. Aligned rectangular few-layer graphene domains on copper surfaces. Chem Mater 2011;23:4543–7. Zhang Y, Zhang L, Kim P, Ge M, Li Z, Zhou C. Vapor trapping growth of single-crystalline graphene flowers: synthesis, morphology, and electronic properties. Nano Lett 2012;12:2810–6. Cho J, Gao L, Tian J, Cao H, Wu W, Yu Q, et al. Atomic-scale investigation of graphene grown on Cu foil and the effects of thermal annealing. ACS Nano 2011;5:3607–13. Wood JD, Schmucker SW, Lyons AS, Pop E, Lyding JW. Effects of polycrystalline Cu substrate on graphene growth by chemical vapor deposition. Nano Lett 2011;11:4547–54. Zhao P, Kumamoto A, Kim S, Chen X, Hou B, Chiashi S, et al. Self-limiting chemical vapor deposition growth of monolayer graphene from ethanol. J Phys Chem C 2013;117:10755–63. Wood JD, Schmucker SW, Haasch RT, Doidge GP, Nienhaus L, Damhorst GL, et al. Improved graphene growth and fluorination on Cu with clean transfer to surfaces. 12th IEEE Conference on Nanotechnology. Birmingham, UK: IEEE; 2012. p. 1–4. Ogawa Y, Hu B, Orofeo CM, Tsuji M, Ikeda KI, Mizuno S, et al. Domain structure and boundary in single-layer graphene grown on Cu (1 1 1) and Cu (1 0 0) films. J Phys Chem Lett 2012;3:219–26. Hu B, Ago H, Ito Y, Kawahara K, Tsuji M, Magome E, et al. Epitaxial growth of large-area single-layer graphene over Cu (1 1 1)/sapphire by atmospheric pressure CVD. Carbon 2012;50:57–65. Orofeo CM, Hibino H, Kawahara K, Ogawa Y, Tsuji M, Ikeda KI, et al. Influence of Cu metal on the domain structure and carrier mobility in single-layer graphene. Carbon 2012;50:2189–96. Takahashi K, Yamada K, Kato H, Hibino H, Homma Y. In situ scanning electron microscopy of graphene growth on polycrystalline Ni substrate. Surf Sci 2012;606:728–32. Ravani F, Papagelis K, Dracopoulos V, Parthenios J, Dassios KG, Siokou A, et al. Graphene production by dissociation of camphor molecules on nickel substrate. Thin Solid Films 2013;527:31–7.

CARBON

9 4 (2 0 1 5) 1 6 0–17 3

[46] Chae SJ, Gu¨nesß F, Kim KK, Kim ES, Han GH, Kim SM, et al. Synthesis of large-area graphene layers on poly-nickel substrate by chemical vapor deposition: wrinkle formation. Adv Mater 2009;21:2328–33. [47] Gao J, Yip J, Zhao J, Yakobson BI, Ding F. Graphene nucleation on transition metal surface. Structure transformation and role of the metal step edge. J Am Chem Soc 2011;133:5009–15. [48] Luo J, Tian P, Pan CT, Robertson AW, Warner JH, Hill EW, et al. Ultralow secondary electron emission of graphene. ACS Nano 2011;5:1047–55. [49] Graf D, Molitor F, Ensslin K, Stampfer C, Jungen A, Hierold C, et al. Spatially resolved Raman spectroscopy of single-and few-layer graphene. Nano Lett 2007;7:238–42. [50] Rao R, Podila R, Tsuchikawa R, Katoch J, Tishler D, Rao AM, et al. Effects of layer stacking on the combination Raman modes in graphene. ACS Nano 2011;5:1594–9. [51] Cong C, Yu T, Saito R, Dresselhaus GF, Dresselhaus MS. Second-order overtone and combination Raman modes of graphene layers in the range of 1690–2150 cm1. ACS Nano 2011;5:1600–5. [52] Araujo PT, Mafra DL, Sato K, Saito R, Kong J, Dresselhaus MS. Unraveling the interlayer-related phonon self-energy renormalization in bilayer graphene. Sci Rep 2012;2. 1017-1–6. [53] He R, Chung TF, Delaney C, Keiser C, Jauregui LA, Shand PM, et al. Observation of low energy Raman modes in twisted bilayer graphene. Nano Lett 2013;13:3594–601. [54] Niilisk A, Kahro T, Kiisk V, Ra¨hn M, Alles H, Aarik J, et al. Raman modes in transferred bilayer CVD graphene. Open Phys 2015;13:34–40. [55] Lu CC, Lin YC, Liu Z, Yeh CH, Suenaga K, Chiu PW. Twisting bilayer graphene superlattices. ACS Nano 2013;7:2587–94. [56] Kim K, Coh S, Tan LZ, Regan W, Yuk JM, Chatterjee E, et al. Raman spectroscopy study of rotated double-layer graphene: misorientation-angle dependence of electronic structure. Phys Rev Lett 2012;108. 246103-1–6. [57] Wang Y, Su Z, Wu W, Nie S, Lu X, Wang H, et al. Four-fold Raman enhancement of 2D band in twisted bilayer graphene: evidence for a doubly degenerate Dirac band and quantum interference. Nanotechnology 2014;25. 335201-1–7. [58] Latil S, Meunier V, Henrard L. Massless fermions in multilayer graphitic systems with misoriented layers: Ab initio calculations and experimental fingerprints. Phys Rev B 2007;76. 201402-1–4.

173

[59] Tiberj A, Camara N, Godignon P, Camassel J. Micro-Raman and micro-transmission imaging of epitaxial graphene grown on the Si and C faces of 6H-SiC. Nanoscale Res Lett 2011;6:1–9. [60] Hass J, Varchon F, Millan-Otoya JE, Sprinkle M, Sharma N, de Heer WA, et al. Why multilayer graphene on 4H-SiC (0 0 0 I) behaves like a single sheet of graphene. Phys Rev Lett 2008;100. 125504-1–4. [61] Carozo V, Almeida CM, Ferreira EH, Cancado LG, Achete CA, Jorio A. Raman signature of graphene superlattices. Nano Lett 2011;11:4527–34. [62] Carozo V, Almeida CM, Fragneaud B, Bedeˆ PM, Moutinho MVO, Ribeiro-Soares J, et al. Resonance effects on the Raman spectra of graphene superlattices. Phys Rev B 2013;88. 085401-1–10. [63] Wu J, Xu H, Mu W, Xie L, Ling X, Kong J, et al. Observation of low-frequency combination and overtone Raman modes in misoriented graphene. J Phys Chem C 2014;118:3636–43. [64] Cong C, Yu T. Enhanced ultra-low-frequency interlayer shear modes in folded graphene layers. Nat Commun 2014;5. 47091–7. [65] Sato K, Park JS, Saito R, Cong C, Yu T, Lui CH, et al. Raman spectra of out-of-plane phonons in bilayer graphene. Phys Rev B 2011;84. 035419-1–5. [66] Campos-Delgado J, Canc¸ado LG, Achete CA, Jorio A, Raskin JP. Raman scattering study of the phonon dispersion in twisted bilayer graphene. Nano Res 2013;6:269–74. [67] Luican A, Li G, Reina A, Kong J, Nair RR, Novoselov KS, et al. Single-layer behavior and its breakdown in twisted graphene layers. Phys Rev Lett 2011;106. 26802-1–4. [68] Hernandez YR, Schweitzer S, Kim JS, Patra AK, Englert J, Lieberwirth I, et al. Turbostratic graphitic microstructures: electronically decoupled multilayer graphene devices with robust high charge carrier mobility, 2013, [In Cornell University Library: arXiv:1301.6087; see http://arxiv.org/ftp/ arxiv/papers/1301/1301.6087.pdf]. [69] Sette F, Hashizume T, Comin F, MacDowell AA, Citrin PH. Coverage and chemical dependence of adsorbate-induced bond weakening in metal substrate surfaces. Phys Rev Lett 1988;61:1384–7. [70] Cheng D, Hou M, Moors M, de Bocarme´ TV, Kruse N. Triggering surface nickel diffusion by adsorption of carbon. Chem Phys Lett 2010;492:63–7.