Stranski–Krastanov and Volmer–Weber CVD ... - ACS Publications

Letter pubs.acs.org/NanoLett

Stranski−Krastanov and Volmer−Weber CVD Growth Regimes To Control the Stacking Order in Bilayer Graphene Huy Q. Ta,‡,†,∥ David J. Perello,§ Dinh Loc Duong,§ Gang Hee Han,§ Sandeep Gorantla,⊥ Van Luan Nguyen,§ Alicja Bachmatiuk,‡,# Slava V. Rotkin,∇ Young Hee Lee,*,∥,§ and Mark H. Rümmeli*,†,‡,# †

College of Physics, Optoelectronics and Energy & Collaborative Innovation Center of Suzhou Nano Science and Technology, Soochow University, Suzhou 215006, China ‡ Centre of Polymer and Carbon Materials, Polish Academy of Sciences, M. Curie-Sklodowskiej 34, Zabrze 41-819, Poland § Center for Integrated Nanostructure Physics, Institute for Basic Science, Sungkyunkwan University, Suwon 16419, Republic of Korea ∥ Department of Energy Science, Department of Physics, Sungkyunkwan University, Suwon 16419, Republic of Korea ⊥ Department of Physics, University of Oslo, Blindern, P.O. Box 1048, 0316 Oslo, Norway # IFW Dresden, P.O. Box 270116, D-01171 Dresden, Germany ∇ Department of Physics and Center for Advanced Materials and Nanotechnology, Lehigh University, Bethlehem, Pennsylvania 18015, United States S Supporting Information *

ABSTRACT: Aside from unusual properties of monolayer graphene, bilayer has been shown to have even more interesting physics, in particular allowing bandgap opening with dual gating for proper interlayer symmetry. Such properties, promising for device applications, ignited significant interest in understanding and controlling the growth of bilayer graphene. Here we systematically investigate a broad set of flow rates and relative gas ratio of CH4 to H2 in atmospheric pressure chemical vapor deposition of multilayered graphene. Two very different growth windows are identified. For relatively high CH4 to H2 ratios, graphene growth is relatively rapid with an initial first full layer forming in seconds upon which new graphene flakes nucleate then grow on top of the first layer. The stacking of these flakes versus the initial graphene layer is mostly turbostratic. This growth mode can be likened to Stranski−Krastanov growth. With relatively low CH4 to H2 ratios, growth rates are reduced due to a lower carbon supply rate. In addition bi-, tri-, and few-layer flakes form directly over the Cu substrate as individual islands. Etching studies show that in this growth mode subsequent layers form beneath the first layer presumably through carbon radical intercalation. This growth mode is similar to that found with Volmer−Weber growth and was shown to produce highly oriented AB-stacked materials. These systematic studies provide new insight into bilayer graphene formation and define the synthetic range where gapped bilayer graphene can be reliably produced. KEYWORDS: Bilayer graphene, stacking order, stacking control, hydrogen role, growth mechanism

A

This unusual characteristic of bilayer graphene has attracted considerable attention for fundamental studies and potential applications in digital electronics and photonics. In the case of twisted BLG its electronic properties critically depend upon the twist angle and has is in itself attracting attention in various fields.5−14 Thus, it is important that reliable and viable synthesis routes for both AB and twisted BLG be found. In this vein, chemical

part from the single layer graphene (SLG), stacking graphene sheets, one atop the other, can lead to very different properties such as superlinear energy-momentum relations and allow one expanding the electronic versatility of this two-dimensional carbon materials.1−4 In the very simple structure, two graphene layers can be arranged in an AA, AB, or a twisted configuration; the physical properties of such bilayer graphene (BLG) are correlated with the stacking order and relative twist angle. Bernal AB-stacked BLG is of significant interest for graphene based field-effect transistors (FETs) and optoelectronic applications because of the feasibility to continuously tune its band gap with a vertical electric field. © 2016 American Chemical Society

Received: July 8, 2016 Revised: September 20, 2016 Published: September 29, 2016 6403

DOI: 10.1021/acs.nanolett.6b02826 Nano Lett. 2016, 16, 6403−6410

Letter

Nano Letters

Figure 1. Relative gas flow windows used to investigate the dependence of graphene growth modes with respect to the CH4 partial pressure (estimated by the flow rate of CH4 divided by the total flow rate) versus the total gas flow (a) (total gas flow includes CH4, H2, and a constant flow 1000 SCCM of Ar). The small dots indicated all the measured points. Additional SEM images correspond to those dots which are marked (1) to (8) are available in Figure S1. (b−d) A set of SEM images showing SK-like bilayer graphene growth corresponding to (b−d) red square spots in (a). The graphene flakes irregular in shape in the SK-like mode. (e−g) A set of SEM images showing VW-like bilayer graphene growth corresponding to (e− g) blue triangle spots in (a). The graphene flakes are regular in shape in the VW-like mode. All scale bars are 2 μm.

second layer grows atop the first. However, studies show bilayer graphene can occur below the first formed layer, viz., under growth.28,29 In the case of under growth it is argued carbon radicals (CHx, x < 4) intercalate between the substrate and the first graphene layer. These intercalated species then decompose and form new graphene beneath the first layer. Another under growth route was recently presented by Hao et al. in which the dissolution and diffusion of carbon atoms through a Cu substrate took place to enable bilayer growth. Oxygen impurities on the Cu surface were shown to function as an activator for hydrocarbon molecule dissociation enhancing the carbon feed rate.30 Here we present a systematic study to understand the stacking order of graphene layers over copper’s dependence on the partial CH4 flow with respect to the total gas flow. This was conducted over a broad range of partial CH4 flows with respect to the total gas flow as indicated in Figure 1. The data show that the uniformity of the stacking order and secondary layer formation (above or below the initial layer) are strongly dependent on the relative gas flow ratios as will be discussed in greater detail later. Our CVD grown multilayer graphene samples show two prevalent growth modes; for relatively low CH4/H2 ratios highly uniform AB-stacked graphene layers are obtained while for relatively high CH4/H2 ratio turbostratic few-layer graphene is formed. These two modes are akin to the Stranski−Krastanov (SK) (high CH4/H2 ratios) and Volmer− Weber (VW) (low CH4/H2 ratios) growth modes which are

vapor deposition (CVD) methods are emerging as promising for large-area, high-quality graphene and few-layer synthetic graphene, usually over transition metals.15−21 Indeed, numerous CVD approaches have been reported. For instance, Ni which has a relatively high carbon solubility22,23 (∼1.3 atomic % at 1000 °C) tends to yield multilayer graphene with various thickness and random staking order.23,24 Multilayer graphene with controllable layer numbers has been reported when using a Ni−Cu alloy substrate; however, homogeneous thickness throughout the sample is limited due to challenges in controlling the uniformity of alloy substrate.22,25−27 Moreover, the stacking order of bilayer graphene often consists of a mixture of twisted and AB-stacked bilayer graphene. Another substrate of interest is Cu which has a low carbon solubility and recent studies show one can obtain AB-stacked bilayer graphene.17−19 However, a key drawback remains, namely, homogeneous stacking order and coverage of bilayer have not as yet been demonstrated. This is in part due to an incomplete understanding on the growth of bilayer graphene. Thus, far, only a limited number of reports seek to explain mechanism of bilayer graphene growth over copper. Two routes gave apparently been identified: Bilayer graphene over growth or epitaxy layer-by-layer growth,18,19 in which carbon radicals or fragments generated from decomposing CH4 originating over some copper catalyst further upstream can deposit further downstream and allow epitaxial growth of a second graphene layer over the first previously formed layer. In other words, the 6404

DOI: 10.1021/acs.nanolett.6b02826 Nano Lett. 2016, 16, 6403−6410

Letter

Nano Letters

Figure 2. Timeline evolution of multilayer graphene growth for the two different growth modes. (a−d) SEM image of time dependence in SK-like growth. (e−h) SEM image of time dependence in VW-like growth. (i) Schematic illustration of the APCVD graphene growth on copper in Stranski−Krastanov and Volmer−Weber growth modes. All scale bars = 1 μm.

two of the three primary modes found in thin film epitaxial growth at crystal surfaces. The SK mode follows a two-step process: initially, complete films of adsorbates form up to a few monolayers thick. Once exceeding a critical layer thickness (which depends on strain and the chemical potential of the deposited film) growth continues through nucleation and coalescence of adsorbate “islands” over the crystal substrate. In the VW mode, growth occurs because the atoms of the deposited material are more strongly bound to each other than they are to the substrate. This leads to island growth before the initial film has fully formed. Moreover, the secondary layers (island growth) form beneath the initial layer. These two modes are schematically represented in Figure 1. The figure also shows clearly the two synthesis windows in which we are able to switch between AB-stacked and twisted multilayer graphene. Moreover, we are able to achieve ca. 95% of ABstacked of bi-, tri-, and multilayer graphene when operating in the VW synthesis mode. Results and Discussion. As mentioned above, the dependence of the CH4/total gas ratio on graphene formation over polished Cu using atmospheric pressure CVD (APCVD) was investigated. Two distinct growth windows in terms of the partial CH4 flow with respect to the total gas flow were found. These are shaded in light orange (relatively high gas flow rates in which SK-like growth occurs) and light blue (relatively low gas flow rates where VW-like growth is obtained) in Figure 1. Scanning electron microscopy (SEM) evaluations of the graphene flakes formed over the Cu substrates show two types of graphene flakes form. In the SK-like growth region (light orange) flakes with various lobes are formed. These flakes, as indicated by Raman spectroscopic studies (discussed later), reside over an initially formed complete layer of graphene, and the flakes residing on the graphene surface have various lobes emanating from them. The SEM data also show creases from the underlying homogeneous graphene layer (see SEM in Figure 1b−d). In the VW-like growth region (light

blue), the graphene flakes tend to be well-faceted. The layers apparently stack in a pyramid-like formation. No initial full coverage of graphene is observed (see SEM in Figure 1e−g). Additional SEM images spanning both the growth windows for SK and VW modes as marked in Figure 1a are provided in Figure S1 in the Supporting Information. The differing characteristics can be observed easily in the SEM images provided in Figure 2. In order to study the evolution timeline of these two growth modes, SEM micrographs were taken at various growth times. Growth rates are significantly different between the two modes. Figure 2a−h shows SEM images of as-produced few-layer graphene grown from each of the two growth modes (SK-like growth and VW-like growth, respectively). The size of the flakes increases with time as indicated in Figure 2 and Figure S2. The average size of the flakes residing of the initially homogeneous graphene layer in SK-like growth range from 3.7 ± 0.2 (μm) to near full (ca. 90%) coverage for growth periods of 3−15 s. The average flake sizes obtained in the VW-like growth mode vary from 0.5 ± 0.1 (μm) to 4.8 ± 0.3 (μm) for much longer growth periods of 20−120 s; namely, VW-like growth is slower than SK-like growth. When looking at the flakes (for both growth modes) in more detail (e.g., Figure 2) one can see changes in contrast (darker regions) which are due to changes in the numbers of layers.31,32 The time data show that graphene growth in both growth modes, in our CVD reactor, is rather rapid. In particular, growth in the SK-like mode is very fast in that after only 3 s an initial full coverage layer of graphene has already formed (confirmed by Raman spectroscopy) and new flakes have also developed. An evaluation of these flakes shows they reside on top of the initial graphene layer which we now elaborate in more detail. Previous reports have shown that graphene grown over Cu(111) faces always form a hexagonal shape and that independent flakes are all oriented with respect to each other such that upon merging no grain boundary forms between the 6405

DOI: 10.1021/acs.nanolett.6b02826 Nano Lett. 2016, 16, 6403−6410

Letter

Nano Letters

Figure 3. Comparison of Raman mapping and spectra revealing the stacking order distribution of graphene flakes from the SK- and VW-like growth modes. (a−c) SEM image, I2D/IG ratio Raman mapping, Raman spectra corresponding to each spot in the Raman mapping image of SK multilayer graphene, respectively. (d−f) SEM image, I2D/IG ratio Raman mapping, Raman spectra corresponding to each spot in the Raman mapping image of VW multilayer graphene, respectively. The scale bar is 2 μm.

defects it becomes active due to symmetry breaking and hence is a useful marker as to the quality of graphene.34,35 The G band arises from the in-plane vibration of sp2 carbon atoms. It is a doubly degenerate (TO and LO) phonon mode (E 2g symmetry) at the Brillouin zone center.35 The 2D band is based on a two phonons lattice vibration process, which unlike the D mode does not activate upon proximity to a defect. Distinct band shape and relative intensity (to the G and 2D band) differences allow the effective evaluation of few-layer graphene and single layer graphene layer number (e.g., Figure S5).17,35 This sensitivity is also true for twisted bilayer graphene in which there exists a relative rotation angle between layers with the relative twist affecting peak position, fwhm, and the relative intensities of the 2D and G modes. These features allow one to broadly classify twisted bilayer graphene into three different types: (i) AB (Bernal) stacked or strongly coupled bilayer, which correlates with a relative rotation angle from 0° to 5°. In this region the Raman spectrum shows that the ratio intensity of 2D peak and G peak is close to 1 and the 2D peak full width at half-maximum (fwhm) increases with respect to that obtained with monolayer graphene. (ii) The G band Raman double resonance mode which correlates with a relative rotation angle range in bilayer graphene from 7° to 14°. In this range, the intensity of G peak is enhanced so that the relative intensity of the 2D band to the G mode is less than 1. (iii) The decoupled or weakly coupled bilayer system occurs for twisted bilayer with large relative rotation angles, from 17° to 30°. This stacking order behaves similar to monolayer graphene in that its 2D to G peak ratio is higher than 1; however, the G mode intensity in weakly coupled bilayer graphene is twice that found for monolayer graphene. In addition, the 2D mode is blueshifted relative to that for monolayer graphene.39−41 Based on these three rotation angle windows, we can better characterize our bilayer samples. The different properties of the few-layer graphene flakes found from SK- and VW-like growth can be evaluated using SEM images matched to Raman spectra mapping. Figure 3a and d shows representative SEM images of SK and VW

domains. This is because the Cu(111) face has strong influence on the graphene because of its small lattice mismatch (4%). In contrast, on non-Cu(111) they found the graphene domains are not aligned.33 With this fact in mind we can carefully examine the flakes formed in our SK-like graphene growth. If our bilayer which was grown under the first layer rests on a Cu(111) face then it must be influenced by Cu substrate leading to the formation of a hexagonal flake which should also be aligned other bilayer flakes nearby on the Cu (111) face. However, in our case on Cu (111) faces we see a mix of hexagonal flakes and flower-like flakes with multiple lobes with different orientations (see Figure S3). This indicates the bilayer flakes are not influenced by the Cu (111) surface and are thus forming on the surface of the initial graphene layer covering the entire substrate. In addition, an EBSD map of the copper substrate (see Figure S4) after the annealing process but before the CVD reaction confirms the presence of Cu(111) faces exist prior to growth. With increasing time, as the flakes grow in size, they merge with each other as can be seen in Figure S2. The merging process is faster for the SK-like growth mode. Moreover, once the flakes have merged, a predominantly bilayer coverage (>95%) exists (Figure 2d), while for the VW-like growth, independent flakes are still clearly observable (e.g., Figure S2h). Even after 120 s the flakes maintaining a pyramid like growth, as is expected in VW growth since atoms are more strongly coupled to each other than the substrate, thus favoring the formation of three-dimensional islands. We now turn to a Raman spectroscopic evaluation of the graphene produced by both the SK- and VW-like growth modes. Raman spectroscopy is a powerful characterization ideal to exploit numerous structural and electronic properties of graphene, such as layer number, stacking order, strain effect, band structure, and doping concentration.34−38 The major features found in the Raman spectrum of graphene are the D band (∼1350 cm−1), the G band (∼1580 cm−1), and the 2D band (∼2670 cm−1). The D band is attributed a ring breathing mode from sp2 carbon rings. The D mode forbidden but near 6406

DOI: 10.1021/acs.nanolett.6b02826 Nano Lett. 2016, 16, 6403−6410

Letter

Nano Letters

Figure 4. HRTEM and SAED characterization of SK- and VW-like grown bilayer graphene. (a) HRTEM image of SK-like grown bilayer graphene showing Moiré pattern, (b) HRTEM image of SK bilayer graphene with a hole allowing edge counting to confirm bilayer, and (c) SAED pattern from SK-like grown bilayer graphene showing two sets of diffraction spots with a 10.56° relative rotation. (d) HRTEM image of VW-like bilayer graphene showing graphene atomic structure from Bernal stacking, (e) HRTEM image of VW-like bilayer graphene with a hole allowing visual confirmation of bilayer graphene, and (f) SAED pattern from AB-stacked VW-like grown bilayer graphene, inset: profile plot of diffraction peak intensities along dashed line. All scale bars are 2 nm.

multilayer graphene transferred on to SiO2 (300 nm)/Si substrate. In panel a, the (SK-like grown) bilayer flakes reside for the most part independently over the initial full coverage graphene layer. Some of the flakes show additional patches of trilayer graphene. In panel d the (VW-like grown) flake consists of various layers as seen by the changes in contrast. The main largest flake is just touch/merging with neighboring flakes. Corresponding Raman maps for the I2D/IG ratio for these equivalent regions are provided in panels b and e, respectively. For the SK-like grown graphene (panel b), the I2D/IG intensity Raman mapping shows inhomogeneous stacking order, including monolayer shaded in light red (I), AB-stacked or strongly coupled bilayer (twist: 0−5°) shaded in dark red (II), G band Raman double resonance bilayer (twist: 7−14°) shaded in dark brown (III), and weakly coupled bilayer (twist: 17− 30°) shaded in orange (IV).39 A more detailed characterization as presented in Figure 3c shows the typical Raman spectra profiles corresponding to each of the shaded regions (I through IV). Monolayer graphene is in pink (region I), AB or near AB stacked bilayer graphene in blue, double resonance bilayer (twist: 7−14°) in red, and weakly coupled bilayer graphene (twist: 17−30°) in green. In contrast with SK-like grown graphene, I2D/IG intensity Raman mapping of VW graphene shows homogeneous AB-stacked order in bilayer, trilayer, and multilayer34,35 regions as shown in Figure 3e and the large-scale Raman mapping in Figure S6. The Raman spectra profiles corresponding to each layer are provided in Figure 3f. To further study of stacking order of flakes grown in the SKand VW-like modes, the samples were transferred on standard lacey TEM grids for high resolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) investigations. (Figure S7 shows the low magnification SEM and TEM images of the SK- and VW-like grown bilayer graphene transferred on lacey TEM grid to highlight the area where HRTEM and SAED were measured.) Figure 4 shows HRTEM micrographs and corresponding SAED data for flakes obtained through SK-like growth (panels a−c) and VW-like

growth (panels d−f). The bilayer graphene shown in panels a and b clearly shows Moiré patterns indicating twisted bilayer graphene, and this is clearly seen by the presence of 12 rather than 6 reflex spots in the SAED pattern from the 3-fold diffraction symmetry for each layer.21 In contrast AB stacked graphene does not show any Moiré pattern as can be seen in panels d and e for the VW-like grown bilayer graphene. In addition, there are only six reflex spots arising from the (10− 10) and (11−20) planes confirming AB stacking. An intensity line profile running across both planes as indicated in the SAED of panel f shows the two inner peaks (10−10) plane are less intense than for the (11−20) planes concomitant with bilayer Bernal graphene.42 In summary, SK-like grown bilayer flakes are often turbostratic (twisted) and VW-like grown few-layer graphene follow AB stacking. Lastly, to confirm the stacking electrically, we examine VWand SK-growth samples in dual-gate transistor devices. A schematic and optical micrograph of an example device with top gate stack (Al2O3/Cr) is detailed in Figure 5a,b. If the stacking is consistently AB stacking, evidence of a bandgap should manifest as a decreased minimum conductance (σsd) as the out-of-plane displacement field is increased. For VWgrowth samples, the conductance displays this phenomenon exactly when plotted as a function of Vtg (while stepping Vbg). σmin follows the prototypical arch-like σmin dependence expected for the existence of a gapped state, as shown in Figure 5c. Conversely, the stacking of samples grown via the SK mechanism is turbostratic. In the case of turbostratic bilayer graphene, this manifests in the dual-gating σsd dependence shown in Figure 5d. No signature of a band gap is observed in the turbostratic case, as expected. On the other hand the BLG conductance in this device is consistent with two weakly interacting graphene layers, which results from a reduced interlayer hopping integral (overlap of A−B atomic sites) as compared to AB stacking. Further examples of turbostratic and SK-type devices (Figure S8), as well as non-AB VW-type 6407

DOI: 10.1021/acs.nanolett.6b02826 Nano Lett. 2016, 16, 6403−6410

Letter

Nano Letters

homogeneous AB stacked growth with well faceted flakes; moreover, as we show below successive or secondary layers grow underneath of the first layer. The primary difference between the two grow modes is the relative amount of hydrogen. In the VW-mode the proportional amount of H2 to CH4 is now very high. The role of hydrogen has been shown to aid CH4 decomposition, but more importantly hydrogen radicals can etch weak carbon bonds and thus control the shape, size and stacking order of the graphene.49 Initially, a first layer nucleates and forms a hexagonal shape. The preferential etching by hydrogen at zigzag or armchair edges favors keeping faceted structures in keeping with a hexagon. Subsequent layers that form below the initial flake also keep their hexagonal shape favoring AB stacking. Given that these additional layers appear in the middle of the primary graphene grain, it is likely their formation occurs at the beginning when the supply of adsorbed active carbon exceeds that which can be used up for growth by the small perimeter of the initial graphene flake.50 In addition, previous studies have shown that, typically for such few-layer islands, secondary layers form beneath the initial graphene layer. This is attributed to the intercalation of carbon radicals between the substrate and the graphene layer just above the substrate.28,29 To determine if our secondary layers (in the VW mode) form beneath or over of the first layer, the graphene etching using laser scanning51 in a scanning confocal Raman spectroscopy system was performed. The results are described in detail in section 3 of the Supporting Information. In short, the laser etching process shows that secondary graphene layers are not affected by the etching laser treatment while the first layer shows severe etching. This shows that the first layer resides on top of the secondary layers and protects them from the etching treatment. In terms of the change in graphene flake morphology, it is not straightforward at this stage to understand the differences, and it is hard to compare since in the SK-like growth the graphene flakes form over graphene, while in the VW-like mode the flakes form over the Cu substrate (under-growth). That said, the formation of hexagon flakes (VW-like growth) is usually attributed to form from the kinetic Wulff reconstruction which is argued to occur by edge-attachment-limited growth,52 while multibranched flakes as often found in the SK-like growth are attributed to diffusion (mass transport)-limited growth.53,54 Conclusion. In this study a broad range of CH4 and H2 flows in APCVD are explored for the formation of bilayer graphene over Cu. Two growth windows with fundamentally different growth outcomes are found. In the first, initially a full coverage of graphene is obtained, and then subsequent graphene flakes nucleate and grow over the initial graphene layer. In the second, bi-, tri-, and few-layer graphene flakes form together over the Cu substrate with secondary flakes forming between the substrate and the initial graphene layer (undergrowth) in keeping with previous studies using a relatively high H partial pressure. These two different growth processes are akin to Stranski−Krastanov and Volmer−Weber growth processes. In addition to the different growth processes, the morphology of the flakes as well as their stacking vary between the two growth windows. Flakes grown in the SK-like growth mode tend to have multiple lobes and have turbostratic stacking, while those flakes formed in the VW-like window are well-faceted and are more hexagonal in shape and have Bernal (AB) stacking. The data provide new insight into the growth of bilayer and few-layer graphene in APCVD.

Figure 5. (a) Schematic of dual-gated BLG transistors. (b) Optical image of VW-growth graphene flake and corresponding dual-gated device. Black dotted lines denote the position of the channel. (c) Conductance data from VW-growth sample as a function of both gate potentials. Data are from S1 in panel b. The minimum conductance trend indicates the formation of a gapped state at large displacement field. (d) SK-growth sample conductance as a function of gate potential, with the conductance minimum displaying a stretched Ushape trend.

devices (Figure S9), can be found in the Supporting Information. Discussion. We now turn to a discussion of the data and the two different growth modes. As can be seen in Figure 1 the SKlike growth region dominates inside the parameter region we investigated, and in this dominant SK-growth zone one can say the relative flow of CH4 to H2 is high. This implies a lot of C species are available as feedstock material, and this is reflected in the very rapid growth of graphene in this region. Indeed, after only 3 s a full layer of graphene has already covered the Cu foil, and new graphene flakes have started to form on the surface of the initial full coverage graphene layer. Moreover 95% bilayer coverage is obtained after 20 s. Although 20 s for near full bilayer coverage is fast as compared to growth in the VW-like mode, relative to the growth rate of the first full graphene layer (less than 3 s) it is significantly slower. This is attributed to the reduced catalytic effect from the Cu foil since it is now covered by the initial graphene layer. Hence, the importance of thermal decomposition is crucial. To demonstrate the relevance of the thermal decomposition of CH4 with H 2 under our experimental conditions we conducted thermodynamic calculations using the NASA computer program CEA43 (Chemical Equilibrium with Applications). Details of the calculations and results are provided in section 2 of the Supporting Information. In short, our calculations confirm the decomposition of CH4 under our experimental conditions (and also at lower temperatures). Our data are in agreement with studies by various other workers.44−48 Indeed, Abánades et al.44 experimentally observed methane thermal decomposition (without a catalyst) already starts from 875 °C, and the decomposition becomes more efficient as the reaction temperature increases, in agreement with our thermodynamic calculations. In our case, Cu may also play a role, because Cu deposits on the quartz reactor tube are found one each side of the Cu foil position (see Figure S10). This Cu deposition, even behind the foil (upstream), is attributed to turbulence in the gas flow. In the VW-like growth zone, growth is fundamentally different in that it is not only slower, but we also obtain near 6408

DOI: 10.1021/acs.nanolett.6b02826 Nano Lett. 2016, 16, 6403−6410

Letter

Nano Letters Methods. CVD Growth. Multilayer graphene was grown over polished copper using atmospheric pressure CVD using methane (99.999%) as the feedstock. In the APCVD system, a horizontal quartz tube of length 75 cm and diameter of 5 cm was used as the reactor chamber. The heater was designed with two separate zones that together form a 60 cm heating zone. Prior to growth, a 100 μm thick copper foil (from Nilaco, 99.96%) was annealed at 1060 °C with 1000 sccm Ar and 200 sccm H2 for 2h, then polished using a chemical mechanical polishing method.55 The polished copper was then mounted in a 2 in. quartz tube chamber and heated up to 1060 °C in 40 min with constant flows of 1000 sccm Ar and 200 sccm H2. After reaching 1060 °C, the sample was annealed for 1 h without changing the gas flow. For the APCVD growth of graphene reactions, times between 3 and 120 s were applied (the growth time was controlled by measuring time between switching on and off the CH4 and H2 mass flow controllers (MFCs)). The gas flow rates explored were CH4 from 1 to 90 sccm and H2: 50 to 350 sccm (Ar 1000 sccm). After the reaction, CH4 and H2 were shut down, and the samples was cooled to room temperature in Ar and then removed. Graphene Transfer. In order to protect the graphene film during Cu etching, a polymethylmethacrylate (PMMA) solution (950 k C4) was spin-coated on the graphene/Cu at 1000 rpm for 60s. To etch the Cu foil, the sample was floated in copper etchant (CE-100, Transene) for ∼30 min. After rinsing in deionized water for a few times, the PMMA/graphene layer was fished onto SiO2 (300 nm)/Si wafer, and PMMA was removed by acetone later.55 The same process was also used for a transfer onto a Cu grid for TEM observations. Characterization. For Raman spectroscopy and mapping, we used a confocal Raman CRM 200 (Witec, Germany) with 100× lens (Olympus, N.A. 0.9) and ∼1 mW power from 532 nm excitation laser. SEM images were obtained with a field emission scanning microscope JSM-7600F (JEOL, Korea) at 15 kV. Transmission electron microscopy (TEM) images and electron diffraction patterns were acquired using a FEI Titan cubed image Cs corrected system at 80 kV.

■

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Institute for Basic Science (IBS-R011-D1) and in part by BK21-plus through the Ministry of Education, Korea. A.B. thanks the National Science Centre for the financial support within the frames of the Sonata Program (Grant agreement 2014/13/D/ST5/02853). M.H.R. acknowledges the DFG (DFG RU1540/15-2) and the National Science Foundation China (Project 51672181).

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.nanolett.6b02826. OM, SEM, EBSD, Raman spectra, Raman mapping, transport data, photographs in Figures S1−S10, thermodynamic calculations, and secondary graphene layer position determination sections (PDF)

■

REFERENCES

(1) Zhang, Y.; Tang, T. T.; Girit, C.; Hao, Z.; Martin, M. C.; Zettl, A.; Crommie, M. F.; Shen, Y. R.; Wang, F. Nature 2009, 459, 820− 823. (2) McCann, E.; Abergel, D. S. L.; Fal’ko, V. I. Solid State Commun. 2007, 143, 110−115. (3) He, W. Y.; Chu, Z. D.; He, L. Phys. Rev. Lett. 2013, 111, 066803. (4) Shallcross, S.; Sharma, S.; Kandelaki, E.; Pankratov, O. A. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 165105. (5) Lopes dos Santos, J. M. B.; Peres, N. M. R.; Castro Neto, A. H. Phys. Rev. Lett. 2007, 99, 256802. (6) Trambly de Laissardière, G.; Mayou, D.; Magaud, L. Nano Lett. 2010, 10, 804−808. (7) Mele, E. J. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 161405. (8) Schmidt, H.; Lüdtke, T.; Barthold, P.; McCann, E.; Fal’ko, V. I.; Haug, R. J. Appl. Phys. Lett. 2008, 93, 172108. (9) Kim, K. S.; Walter, A. L.; Moreschini, L.; Seyller, T.; Horn, K.; Rotenberg, E.; Bostwick, A. Nat. Mater. 2013, 12, 887−892. (10) Hicks, J.; Sprinkle, M.; Shepperd, K.; Wang, F. Phys. Rev. B 2011, 83, 205403. (11) Li, G.; Luican, A.; Lopes dos Santos, J. M. B.; Castro Neto, A. H.; Reina, A.; Kong, J.; Andrei, E. Y. Nat. Phys. 2009, 6, 109−113. (12) Luican, A.; Li, G.; Reina, A.; Kong, J.; Nair, R. R.; Novoselov, K. S.; Geim, A. K.; Andrei, E. Y. Phys. Rev. Lett. 2011, 106, 126802. (13) Wang, Y.; Ni, Z.; Liu, L.; Liu, Y.; Cong, C.; Yu, T.; Wang, X.; Shen, D.; Shen, Z. ACS Nano 2010, 4, 4074−4080. (14) Tan, Z.; Yin, J.; Chen, C.; Wang, H.; Lin, L.; Sun, L.; Wu, J.; Sun, X.; Yang, H.; Chen, Y.; Peng, H.; Liu, Z. ACS Nano 2016, 10, 6725−6730. (15) Yao, Y.; Li, Z.; Lin, Z.; Moon, K. S.; Agar, J.; Wong, C. J. Phys. Chem. C 2011, 115, 5232−5238. (16) Lee, S.; Lee, K.; Zhong, Z. Nano Lett. 2010, 10, 4702−4707. (17) Luo, Z.; Yu, T.; Shang, J.; Wang, Y.; Lim, S.; Liu, L.; Gurzadyan, G. G.; Shen, Z.; Lin, J. Adv. Funct. Mater. 2011, 21, 911−917. (18) Yan, K.; Peng, H.; Zhou, Y.; Li, H.; Liu, Z. Nano Lett. 2011, 11, 1106−1110. (19) Liu, L.; Zhou, H.; Cheng, R.; Yu, W. J.; Liu, Y.; Chen, Y.; Shaw, J.; Zhong, X.; Huang, Y.; Duan, X. ACS Nano 2012, 6, 8241−8249. (20) Sun, Z.; Raji, A. R. O.; Zhu, Y.; Xiang, C.; Yan, Z.; Kittrell, C.; Samuel, E. L. G.; Tour, J. M. ACS Nano 2012, 6, 9790−9796. (21) Lu, C. C.; Lin, Y. C.; Liu, Z.; Yeh, C. H.; Suenaga, K.; Chiu, P. W. ACS Nano 2013, 7, 2587−2594. (22) Wu, Y.; Chou, H.; Ji, H.; Wu, Q.; Chen, S.; Jiang, W.; Hao, Y.; Kang, J.; Ren, Y.; Piner, R. D.; Ruoff, R. S. ACS Nano 2012, 6, 7731− 7738. (23) Li, X.; Cai, W.; Colombo, L.; Ruoff, R. S. Nano Lett. 2009, 9, 4268−4272. (24) Chae, S. J.; Güneş, F.; Kim, K. K.; Kim, E. S.; Han, G. H.; Kim, S. M.; Shin, H. J.; Yoon, S. M.; Choi, J. Y.; Park, M. H.; Yang, C. W.; Pribat, D.; Lee, Y. H. Adv. Mater. 2009, 21, 2328−2333. (25) Chen, S.; Cai, W.; Piner, R. D.; Suk, J. W.; Wu, Y.; Ren, Y.; Kang, J.; Ruoff, R. S. Nano Lett. 2011, 11, 3519−3525. (26) Yang, C.; Wu, T.; Wang, H.; Zhang, G.; Sun, J.; Lu, G.; Niu, T.; Li, A.; Xie, X.; Jiang, M. Small 2016, 12, 2009−2013.

AUTHOR INFORMATION

Corresponding Authors

*E-mail address: [email protected]. *E-mail address: [email protected]. Author Contributions

Y.-H.L. and M.H.R. conceived the project. H.Q.T., D.J.P., D.L.D., A.B., and S.D. performed the experiments. H.Q.T. prepared the samples. H.Q.T., D.L.D., and G.-H.H. conducted the SEM studies and Raman characterization. D.J.P. did the devices fabrication and electronic transport characterization. A.B. and S.D. completed the TEM measurements. All authors participated in the data analysis and manuscript preparation. 6409

DOI: 10.1021/acs.nanolett.6b02826 Nano Lett. 2016, 16, 6403−6410

Letter

Nano Letters (27) Takesaki, Y.; Kawahara, K.; Hibino, H.; Okada, S.; Tsuji, M.; Ago, H. Chem. Mater. 2016, 28, 4583−4592. (28) Li, Q.; Chou, H.; Zhong, J. H.; Liu, J. Y.; Dolocan, A.; Zhang, J.; Zhou, Y.; Ruoff, R. S.; Chen, S.; Cai, W. Nano Lett. 2013, 13, 486− 490. (29) Nie, S.; Wu, W.; Xing, S.; Yu, Q.; Bao, J.; Pei, S.; McCarty, K. F. New J. Phys. 2012, 14, 093028. (30) Hao, Y.; Wang, L.; Liu, Y.; Chen, H.; Wang, X.; Tan, C.; Nie, S.; Suk, J. W.; Jiang, T.; Liang, T.; Xiao, J.; Ye, W.; Dean, C. R.; Yakobson, B. I.; McCarty, K. F.; Kim, P.; Hone, J.; Colombo, L.; Ruoff, R. S. Nat. Nanotechnol. 2016, 11, 426−431. (31) Xie, J.; Spallas, J. Microsc. Microanal. 2013, 19, 370−371. (32) Park, M. H.; Kim, T. H.; Yang, C. W. Surf. Interface Anal. 2012, 44, 1538−1541. (33) Nguyen, V. L.; Shin, B. G.; Duong, D. L.; Kim, S. T.; Perello, D.; Lim, Y. J.; Yuan, Q. H.; Ding, F.; Jeong, H. Y.; Shin, H. S.; Lee, S. M.; Chae, S. H.; Vu, Q. A.; Lee, S. H.; Lee, Y. H. Adv. Mater. 2015, 27, 1376−1382. (34) Malard, L. M.; Pimenta, M. A.; Dresselhaus, G.; Dresselhaus, M. S. Phys. Rep. 2009, 473, 51−87. (35) Ni, Z.; Wang, Y.; Yu, T.; Shen, Z. Nano Res. 2008, 1, 273−291. (36) Lee, J. E.; Ahn, G.; Shim, J.; Lee, Y. S.; Ryu, S. Nat. Commun. 2012, 3, 1024. (37) Neumann, C.; Reichardt, S.; Venezuela, P.; Drögeler, M.; Banszerus, L.; Schmitz, M.; Watanabe, K.; Taniguchi, T.; Mauri, F.; Beschoten, B.; Rotkin, S. V.; Stampfer, C. Nat. Commun. 2015, 6, 8429. (38) Neumann, C.; Reichardt, S.; Drögeler, M.; Terrés, B.; Watanabe, K.; Taniguchi, T.; Beschoten, B.; Rotkin, S. V.; Stampfer, C. Nano Lett. 2015, 15, 1547−1552. (39) Kim, K.; Coh, S.; Tan, L. Z.; Regan, W.; Yuk, J. M.; Chatterjee, E.; Crommie, M. F.; Cohen, M. L.; Louie, S. G.; Zettl, A. Phys. Rev. Lett. 2012, 108, 246103. (40) Ni, Z.; Liu, L.; Wang, Y.; Zheng, Z.; Li, L. J.; Yu, T.; Shen, Z. Phys. Rev. B: Condens. Matter Mater. Phys. 2009, 80, 125404. (41) Havener, R. W.; Zhuang, H.; Brown, L.; Hennig, R. G.; Park, J. Nano Lett. 2012, 12, 3162−3167. (42) Warner, J. H.; Schaeffel, F.; Bachmatiuk, A.; Rummeli, M. H. Graphene: fundamentals and emergent applications; Elsevier: Waltham, 2013; p 282. (43) https://www.grc.nasa.gov/WWW/CEAWeb/ceaHome.htm. (44) Abánades, A.; Ruiz, E.; Ferruelo, E. M.; Hernández, F.; Cabanillas, A.; Martínez-Val, J. M.; Rubio, J. A.; López, C.; Gavela, R.; Barrera, G.; Rubbia, C.; Salmieri, D.; Rodilla, E.; Gutiérrez, D. Int. J. Hydrogen Energy 2011, 36, 12877−12886. (45) Hirsch, D.; Epstein, M.; Steinfeld, A. Int. J. Hydrogen Energy 2001, 26, 1023−1033. (46) Dahl, J. K.; Buecher, K. J.; Weimer, A. W.; Lewandowski, A.; Bingham, C. Int. J. Hydrogen Energy 2004, 29, 725−736. (47) Abánades, S.; Flamant, G. Sol. Energy 2006, 80, 1321−1332. (48) Kogan, M.; Kogan, A. Int. J. Hydrogen Energy 2003, 28, 1187− 1198. (49) Vlassiouk, I.; Regmi, M.; Fulvio, P.; Dai, S.; Datskos, P.; Eres, G.; Smirnov, S. ACS Nano 2011, 5, 6069−6076. (50) Robertson, A. W.; Warner, J. H. Nano Lett. 2011, 11, 1182− 1189. (51) Han, G. H.; Chae, S. J.; Kim, E. S.; Güneş, F.; Lee, I. H.; Lee, S. W.; Lee, S. Y.; Lim, S. C.; Jeong, H. K.; Jeong, M. S.; Lee, Y. H. ACS Nano 2011, 5, 263−268. (52) Artyukhov, V. I.; Liu, Y.; Yakobson, B. I. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 15136−15140. (53) Witten, T. A., Jr.; Sander, L. M. Phys. Rev. Lett. 1981, 47, 1400− 1403. (54) Bartelt, N. C.; McCarty, K. F. MRS Bull. 2012, 37, 1158−1165. (55) Han, G. H.; Güneş, F.; Bae, J. J.; Kim, E. S.; Chae, S. J.; Shin, H. J.; Choi, J. Y.; Pribat, D.; Lee, Y. H. Nano Lett. 2011, 11, 4144−4148.

6410

DOI: 10.1021/acs.nanolett.6b02826 Nano Lett. 2016, 16, 6403−6410