Probing graphene grain boundaries with optical microscopy
LETTER doi:10.1038/nature11562 Probing graphene grain boundaries with optical microscopy Dinh Loc Duong1,2*, Gang Hee Han2*, Seung Mi Lee3, Fethullah Gunes2, Eun Sung Kim2, Sung Tae Kim2, Heetae Kim2, Quang Huy Ta2, Kang Pyo So2, Seok Jun Yoon2, Seung Jin Chae1,2, Young Woo Jo2, Min Ho Park4, Sang Hoon Chae2, Seong Chu Lim2, Jae Young Choi2,5 & Young Hee Lee1,2 Grain boundaries in graphene are formed by the joining of islands during the initial growth stage, and these boundaries govern transport properties and related device performance1,2 Although information on the atomic rearrangement at graphene grain boundaries can be obtained using transmission electron microscopy3,4 and scanning tunnelling microscopy2,5–8, large-scale information regarding the distribution of graphene grain boundaries is not easily accessible Here we use optical microscopy to observe the grain boundaries of large-area graphene (grown on copper foil) directly, without transfer of the graphene This imaging technique was realized by selectively oxidizing the underlying copper foil through graphene grain boundaries functionalized with O and OH radicals generated by ultraviolet irradiation under moisture-rich ambient conditions: selective diffusion of oxygen radicals through OH-functionalized defect sites was demonstrated by density functional calculations The sheet resistance of large-area graphene decreased as the graphene grain sizes increased, but no strong correlation with the grain size of the copper was revealed, in contrast to a previous report9 Furthermore, the influence of graphene grain boundaries on crack propagation (initialized by bending) and termination was clearly visualized using our technique Our approach can be used as a simple protocol for evaluating the grain boundaries of other a two-dimensional layered structures, such as boron nitride and exfoliated clays Optical microscopy is an important tool for characterizing graphene at large scales Although graphene is only one atom thick, a given area can have more than one layer growing on it; areas with differing numbers of layers can be distinguished by their contrast difference10 Optical birefringence from transferred graphene covered by a liquid crystal has been used to visualize graphene grain boundaries (GGBs), but the grain boundaries of the copper substrate were visualized, and looked like GGBs9 This last observation contradicts previous reports based on scanning electron microscopy (SEM) in which GGBs and copper grain boundaries were found to be independent11 Direct characterization of GGBs on copper foil is required Our work examines graphene grown directly on copper without the need for a transfer process, which often leads to undesired artefacts, such as additional traces from copper substrates, wrinkles and/or cracks12 The primary objective of this Letter is to visualize grain boundaries in large-area graphene, and distinguish them from grain boundaries of the copper substrate, using an optical microscope This approach is based on the robust oxidization of copper foil at room temperature via the selective diffusion of O and OH radicals through the GGBs (Fig 1a) These radicals are generated using ultraviolet irradiation c b 10 μm hν O2, H2O hν O•, OH• Graphene grain boundary O2 OH• H2O 10 μm d Oxidized e Graphene grain boundary Oxidized copper μm Figure | Observation of graphene grain boundaries (GGBs) after ultraviolet exposure under moisture-rich ambient conditions a, Diagram of the ultraviolet treatment of a graphene/Cu sample The copper under the GGBs was oxidized by radicals; the lines of oxidized copper were broadened during continuing oxidation and thereby became visible using an optical microscope b, c, Optical images of graphene/Cu before (b) and after (c) oxidation The GGBs were visible after oxidation d, SEM image of oxidized graphene/Cu The GGBs (coincident with oxidized copper seen as white dotted lines) intersected the Cu grain boundary (between the dark central region and the surrounding grey region) e, AFM force image of the position marked by a red square in c Sungkyunkwan Advanced Institute of Nanotechnology, Sungkyunkwan University, Suwon 440-746, South Korea 2Department of Energy Science, BK21 Physics Division, Graphene Center, Sungkyunkwan University, Suwon 440-746, South Korea 3Center for Nanocharacterization, Korea Research Institute of Standards and Science, Daejeon 305-340, South Korea 4School of Advanced Materials Science and Engineering, Sungkyunkwan University, Suwon 440-746, South Korea 5Graphene Center at Samsung Advanced Institute of Technology (SAIT), PO Box 111, Suwon 440-600, South Korea *These authors contributed equally to this work 1 O C T O B E R 2 | VO L | N AT U R E | ©2012 Macmillan Publishers Limited All rights reserved RESEARCH LETTER G-band (1,590 to 1,597 cm21) and the G9-band (2,700 to 2,717 cm21) in addition to the D-band development inside the grain after ultraviolet treatment is indicative of graphene oxidation13,14 Interestingly, at the grain boundary, both the G-band and the G9-band peaks were redshifted compared with those within the grain, respectively from 1,597 to 1,594 cm21 and from 2,717 to 2,688 cm21 This redshift may be explained as follows: in addition to the oxidation of the pristine graphene at the grain boundary, the oxidized copper formed under the GGB has a greater volume than the pristine copper, and this increased volume forced the uppermost graphene to experience a tensile strain15,16 The mappings in D-band, G-band and G9-band (Fig 2c–e) matched well with the GGBs observed using optical microscopy We emphasize here that GGBs cannot be observed using confocal Raman mapping of graphene unless prior ultraviolet treatment has been applied (Supplementary Fig 4) The observation of GGBs using optical microscopy after ultraviolet treatment is possible owing to the morphological changes in the sample Three types of line profile were detected using topological AFM before and after ultraviolet irradiation (Fig 3a): (1) the GGBs; (2) a wrinkle caused by a GGB; and (3) a Cu grain boundary The step-like GGBs (with a step height of 2.5 nm) were visualized more clearly after ultraviolet treatment (Fig 3c top), which enhanced the step height to approximately 15 nm because of the expanded volume of oxidized Cu at the boundary The width of the oxidized boundary increased to approximately 500–600 nm, which is sufficiently large to be observed using an optical microscope The second type of GGB line overlapped the wrinkle line and was approximately nm high12,17 This observation is consistent with a previous theoretical report that predicted wrinkles might be formed from GGBs18 After ultraviolet irradiation, the GGB was distinguishable because of its increased height (Fig 3c middle) This effect is similar to the previous example but the wrinkle line remained unchanged On the other hand, the Cu grain boundary was initially grooved to a depth of 15 nm, and this depth was not altered by oxidation (Fig 3c bottom) This observation supports previous reports of growing graphene sliding over the Cu grain boundary without creating defects11 but contrasts with a report of strongly correlated GGBs and Cu grain boundaries9 Our results provide further evidence of the clear distinction between GGBs and Cu grain boundaries Further analysis of the under moisture-rich ambient conditions The GGBs are very narrow, but the width of the oxidized copper foil beneath the defective GGBs is increased by the continuous supply of oxidizing radicals through the grain boundary In addition, oxidized copper occupies a larger volume than copper, thereby allowing the visualization of the GGBs using an optical microscope (Fig 1a) In general, GGBs are not visible using optical microscopy because of the nanoscale width of the boundaries2–8 Although copper grain boundaries were clearly observed in the optical microscopy images because of their large width, GGBs were not visible (Fig 1b) The CVD-grown large-area graphene on the Cu foil was used directly without transfer in our study (see Methods) After the graphene/Cu sample was irradiated by ultraviolet light under moisture-rich ambient conditions, the GGBs were visible as thin dark lines Some of the GGB lines crossed over the Cu grain boundaries, demonstrating the formation of GGBs regardless of the location of the Cu grain boundaries, which is consistent with previous investigations using SEM11 Similar GGBs were also observed as white lines in scanning electron microscopy (SEM) images (Fig 1d) The copper grain boundaries were clearly distinguishable from the GGBs on the basis of their channelling contrast levels (that is, different orientations of the copper crystals result in different brightnesses; Supplementary Fig 1) It is interesting to note the different thicknesses of the GGB lines, which implies that the nature of the defects in GGBs is complicated Higher oxygen content was observed at the GGBs than within the graphene grains (Supplementary Fig 2) The GGBs in the optical image also matched well with those present in the non-contact force image obtained using tapping-mode atomic force microscopy (AFM; Fig 1e) The GGBs visible in the optical image (Fig 2a) were also confirmed using confocal Raman mapping The CVD-grown graphene/Cu exhibited a large intensity difference between the G9 and G bands (a high G9/ G ratio) with no appreciable D-band (Fig 2b), thereby demonstrating a high-quality monolayer of graphene With ultraviolet treatment, the D-band developed both at the grain boundary and within the grain, but occurred more strongly at the grain boundary This developed D-band intensity was the result of the increased oxygen content, which increased from 2.8 to 14.2 at.%, as confirmed using X-ray photoelectron spectroscopy analysis (Supplementary Fig 3) The blueshift of the Optical image 800 Intensity (a.u.) a b Pristine graphene Oxidized graphene on GB Oxidized graphene in grain 600 400 D G 200 1,200 μm d 1,400 2,688 1,594 1,597 1,590 c D-band intensity G′ 2,700 1,600 2,600 Wavelength shift (cm–1) G-band intensity e 2,717 2,800 G′-band position 200 200 2,720 a.u a.u cm–1 μm Figure | Confocal Raman mapping of GGBs a, b, Optical image (a) and Raman spectra (b) of graphene/Cu The blueshift of the G- and G9-bands in addition to the development of the D-band within the grain provide evidence of graphene oxidation The redshift of the G- and G9-band peaks on the grain μm 2,680 boundary compared with those within the grain was caused by the tensile strain developed in the graphene owing to the expanded underlying oxidized copper c–e, Maps of the intensities of the D-band (c) and G-band (d), and the shift of the G9-band (e) | N AT U R E | VO L | 1 O C TO B E R 2 ©2012 Macmillan Publishers Limited All rights reserved LETTER RESEARCH Before UV a d c 15 1 Wrinkles Height (nm) μm After UV b 200 400 600 800 15 1,000 1,200 Before UV After UV I I 10 μm 10 e III III I 10 3 II Before UV After UV 0 200 400 600 800 1,000 1,200 Before UV After UV –5 –10 μm –15 10 μm 500 1,000 f 1,500 2,000 2,500 Line profile (nm) 3,000 3,500 8.17 Diffusion location Diffusion path (Å) 2.92 O OH Ideal SW 3.4 4.1 OH 4.70 Barrier (eV) OH-functionalized SW 13.60 Diffusion path Figure | Height profiles of various topological GGBs, and the oxidation mechanism a, b, Topological images of various GGBs; (1) GGB, (2) a wrinkle caused by a GGB, and (3) a Cu grain boundary c, Height profiles of the selected positions 1–3 d, e, Optical images of graphene/Cu during early growth (for s) before (d) and after (e) oxidation Three distinct regions are identified: (I) a typical graphene island, (II) a highly defective graphene region, and (III) Cu foil with incomplete graphene growth f, Density functional calculations for diffusion of radicals through a heptagon of OH-functionalized Stone–Wales defects (SW) Stone–Wales defects were enlarged and deactivated by the adsorbed radicals morphology of oxidized copper and graphene after ultraviolet irradiation is shown in Supplementary Figs and A GGB is typically formed by the joining of adjacent graphene islands that were initiated by nucleation seeds To determine the relevance of our approach to observing the initial growth stage of graphene islands, graphene was grown for only s using CVD conditions similar to those described in Methods Three distinct regions near a Cu grain boundary line were observed under these conditions (Fig 3d): a typical graphene island (feature I), a highly defective graphene region (feature II), and Cu foil with incomplete graphene growth (feature III) The black spot in the middle of the grain (Fig 3d) is a nucleation seed that has also previously been observed12 The graphene/Cu shown in Fig 3d was then subjected to ultraviolet irradiation under humid conditions; in Fig 3e we show the area corresponding to Fig 3d.It is interesting to note the occurrence of divergent grain boundary lines within the island centred at the nucleation seed, which is in good agreement with previous confocal Raman mapping and TEM observations6,12 The highly defective graphene region and the graphene incompletely covering the Cu foil were readily oxidized and converted into black spots by the facilitation of radical diffusion Thus, microstructures developed during the initial growth stage can be clearly observed using a simple optical microscope The selective oxidation of Cu foil through the GGBs is dependent on the oxidation conditions For example, the humidity level during ultraviolet irradiation is crucial in developing clear oxidation patterns At low humidity levels (less than 25%), the grain boundaries were only partially visible (Supplementary Fig 7a–d) Even when the oxidation time was prolonged to 30 at these low humidity levels, GGBs were not observed Pure ozone could not oxidize the graphene or the copper substrate at room temperature, as reported previously19,20 The presence of H2O was necessary to create OH radicals under ultraviolet irradiation21,22; these radicals are necessary to functionalize graphene defects to further facilitate copper oxidation (see simulations in Methods) At a humidity level of 25%, the black spots began to appear These black spots, which may be impurities introduced during the preparation of the Cu foil, acted as nucleation seeds12 The black spots could also be attributable to defects in the graphene At higher humidity levels, the size of the black spots became larger, and the GGB lines were clearly visible However, the size of the black spots and the GGB lines did not increase at a greater humidity level (more than 66%) or with longer oxidation times (Supplementary Fig 5e–h) It is important to note that in addition to the oxidation of the graphene (by forming epoxide, hydroxyl and carboxyl groups particularly at the defect sites), the copper foil was also oxidized to give exclusively Cu(OH)2 (Supplementary Fig 8) We performed density functional calculations to understand the oxidation mechanism associated with GGB defects (Fig 3f, Supplementary Figs and 10, Supplementary Table 1) The Stone–Wales defect experimentally observed at GGBs3,4 was modelled The height of the barrier experienced by species diffusing through the graphene was consistently reduced at the functionalized Stone–Wales defects, in particular, those functionalized by OH It was interesting to see that the O radical had a smaller diffusion barrier height than the OH radical In other words, the role of OH radicals was to reduce the diffusion barrier height by functionalizing defects Also, the absolute values of the diffusion barrier height could be lowered by taking into account effects due to the substrate and excited states (see Methods) Although graphene was oxidized and strained by underlying oxidized copper, it showed no cracks (in optical micrographs or SEM images) after transfer (Supplementary Figs 11 and 12) The sheet resistance of transferred graphene also recovered to 426 ohms per square (V %21) after heat treatment at 600 uC in vacuum for h, which was comparable to the 400 V %21 of the pristine graphene (PMMA removed; see Methods for details) The recovery of the electrical properties of 1 O C T O B E R 2 | VO L | N AT U R E | ©2012 Macmillan Publishers Limited All rights reserved RESEARCH LETTER graphene is often useful for studies of GGB-related intrinsic characteristics of graphene To provide proof of concept and to demonstrate the power of our approach, several graphene samples were prepared using different growth temperatures during CVD For each sample, the Cu foil was annealed for 30 min, followed by of graphene growth at a given temperature The range of graphene grain sizes was estimated to be of the order of micrometres on the basis of TEM examination (Supplementary Fig 13) The samples then received an ultraviolet treatment time identical to that described in the experimental section in Methods (10 min) We found that the above-mentioned black spots and GGB lines occurred regardless of the growth temperature (Fig 4a–d) The number of graphene grains in an area 57 mm 76 mm was counted to obtain an average grain size per unit area As the growth temperature increased, the sheet resistance of the graphene gradually decreased to ,300 V %21 for a growth temperature of 1,060 uC, and the average graphene grain size increased to ,72 mm2 (Fig 4e) The copper grain size fluctuated within the range of 10,000–15,000 mm2; however, a correlation between copper grain size and graphene growth temperature was not observed The sheet resistance of the graphene decreased as its grain size increased Other growth conditions with different annealing temperatures (while maintaining the same growth temperatures) were also evaluated but resulted in similar behaviours (Supplementary Fig 14) Thus, increasing the graphene grain size is a key factor in 1,020 °C e oC 1,040 °C c d 10,000 Copper 75 360 60 Resistance 320 45 Graphene 280 30 Grain size (μm) 10 μm 10 μm 15,000 400 f g 360 340 320 300 280 1,000 1,010 1,020 1,030 1,040 1,050 1,060 Temperature (°C) 1,060 °C 380 Sheet resistance (Ω –1) Sheet resistance (Ω –1) b Sheet resistance (Ω –1) 1,000 °C a improving the graphene quality, whereas changing the Cu grain size is not effective in altering the sheet resistance of graphene Conductance AFM measurements showed that different GGBs have different conductances, and these GGBs were found to be consistent with the morphological GGBs observed after ultraviolet exposure (Fig 4g–i and Supplementary Figs 15 and 16) The sheet resistance can be fitted to V Vo[1 (A/Ac)2n], where Vo is the sheet resistance of the intrinsic graphene without grain boundary scattering (or an infinite grain size), A is the average grain size, Ac is a fitting parameter, and n is an exponent We obtained a value of Vo (230 V %21), which to our knowledge has not been estimated previously The theoretical limit of Vo is 30 V %21 (refs 23, 24) In our case, the value is larger owing to scattering from defects and the substrate within the grain This implies that minimizing defects—such as point defects, wrinkles and ripples—and substrate scattering will be important for improving conductivity, in addition to enlarging the graphene grain size To demonstrate our method in a more advanced application, we visualized fracture propagation on bending, using optical microscopy (Fig 4j–m) Fractures propagated preferentially normal to the strain direction and terminated at GGBs (white squares; Fig 4k) As the radius of curvature decreased, more cracks were created and propagated through GGBs, and in some cases, propagation directions were altered at the GGB lines (Fig 4l, m) Our method is not limited to graphene, and can be generalized to analyse the defects and grain size 40 30 h i 50 60 Graphene grain size (μm2) 70 10 μm 10 μm 10 μm Current (nA) μm μm –1 –2 –3 –4 –5 500 1,000 1,500 2,000 2,500 3,000 Line profile (nm) d =16 mm j k d = 10 mm l Crack blockage Crack initiation d = mm m d = mm Crack propagation GB Crack d Tension Bending 10 μm 10 μm Figure | The correlation between GGBs and sheet resistance a–d, Oxidized graphene/Cu grown at indicated temperatures from 1,000 to 1,060 uC e, f, Graphene sheet resistance versus graphene grain size for different growth conditions Error bars, s.d (n 4) In f, a fitted curve for the sheet resistance is arrowed; V V0[1 (A/Ac)2n] (see text for details) g–i, Conductance AFM of graphene/Cu before oxidation, showing two different types of GGBs In i, line scans across small (big) grain boundaries are 10 μm 10 μm indicated by blue (green) curves j–m, Fracture propagation in graphene on Cu as a function of radius of curvature (d/2, see inset in j) Black dots that appeared after oxidation show the position of cracks (red arrows) GGBs are indicated by blue arrows In k and m, the left inset is shown magnified in the right inset GB, grain boundary Fractures propagated normal to the strain direction, terminated (k, l) and changed direction at the GGBs (m) | N AT U R E | VO L | 1 O C TO B E R 2 ©2012 Macmillan Publishers Limited All rights reserved LETTER RESEARCH distribution of other two-dimensional layered structures, such as boron nitride (Supplementary Figs 17 and 18) and exfoliated clays METHODS SUMMARY Ultraviolet oxidization of graphene on copper Graphene on copper foil was placed into a chamber equipped with a low-pressure Hg lamp (LH-arc, Lichtzen, with an output of 20 mW cm22, with the majority of emitted light at a wavelength of 254 nm and approximately 10% of light at a wavelength of 185 nm)13 Humidity was introduced into the chamber by connecting it to a water bubbler The humidity level in the chamber was monitored using a hygro-thermometer (accuracy of 63%) The chamber was continuously ventilated throughout the experiments to maintain a constant pressure After reaching the required humidity level, the water bubbler was disconnected from the chamber The graphene/Cu substrate was then irradiated with ultraviolet light for 10 to oxidize the samples Humid air undergoes the following reactions under ultraviolet irradiation21,22: reactions (1), (2) and (4) require ultraviolet light to proceed, reaction (3) is thermally activated O2 R O3 (1) O3 H2O R O2 H2O2 N (2) 2O3 H2O2 R 2OH 3O2 (3) H2O R HN OHN (4) Full Methods and any associated references are available in the online version of the paper Received 23 May; accepted September 2012 Published online October 2012 10 Blake, P et al Making graphene visible Appl Phys Lett 91, 063124 (2007) 11 Li, X et al Large-area synthesis of high-quality and uniform graphene films on copper foils Science 324, 1312–1314 (2009) 12 Han, G H et al Influence of copper morphology in forming nucleation seeds for graphene growth Nano Lett 11, 41444148 (2011) 13 Guănesá, F et al UV-light-assisted oxidative sp3 hybridization of graphene NANO 6, 409–418 (2011) 14 Jin, Z et al Click chemistry on solution-dispersed graphene and monolayer CVD graphene Chem Mater 23, 3362–3370 (2011) 15 Malarda, L M Pimentaa, M A., Dresselhaus, G & Dresselhaus, M S Raman spectroscopy in graphene Phys Rep 473, 51–87 (2009) 16 Huang, M et al Phonon softening and crystallographic orientation of strained graphene studied by Raman spectroscopy Proc Natl Acad Sci 106, 7304–7308 (2009) 17 Chae, S J et al Synthesis of large-area graphene layers on poly-nickel substrate by chemical vapor deposition: wrinkle formation Adv Mater 21, 2328–2333 (2009) 18 Liu, Y & Yakobson, B I Cones, pringles, and grain boundary landscapes in graphene topology Nano Lett 10, 2178–2183 (2010) 19 Lee, G., Lee, B., Kim, J & Cho, K Ozone adsorption on graphene: ab initio study and experimental validation J Phys Chem C 113, 14225–14229 (2009) 20 Jandhyala, S et al Atomic layer deposition of dielectrics on graphene using reversibly physisorbed ozone ACS Nano 6, 2722–2730 (2012) 21 Feiyan, C., Pehkonen, S O & Ray, M B Kinetics and mechanisms of UV-photodegradation of chlorinated organics in the gas phase Wat Res 36, 4203–4214 (2002) 22 Wang, J H & Ray, M B Application of ultraviolet photooxidation to remove organic pollutants in the gas phase Separ Purif Tech 19, 11–20 (2000) 23 Chen, J.-H., Jang, C., Xiao, S., Ishigami, M & Fuhrer, M S Intrinsic and extrinsic performance limits of graphene devices on SiO2 Nature Nanotechnol 3, 206–209 (2008) 24 Jeong, C., Nair, P., Khan, M., Lundstrom, M & Alam, M A Prospects for nanowiredoped polycrystalline graphene films for ultratransparent, highly conductive electrodes Nano Lett 11, 5020–5025 (2011) Supplementary Information is available in the online version of the paper Yazyev, O V & Louie, S G Electronic transport in polycrystalline graphene Nature Mater 9, 806–809 (2010) Yu, Q et al Control and characterization of individual grains and grain boundaries in graphene grown by chemical vapour deposition Nature Mater 10, 443–449 (2011) Huang, P Y et al Grains and grain boundaries in single-layer graphene atomic patchwork quilts Nature 469, 389–392 (2011) Kim, K et al Grain boundary mapping in polycrystalline graphene ACS Nano 5, 2142–2146 (2011) Tian, J., Cao, H., Wu, W., Yu, Q & Chen, Y P Direct imaging of graphene edges: atomic structure and electronic scattering Nano Lett 11, 3663–3668 (2011) Rasool, H I et al Atomic-scale characterization of graphene grown on copper (100) single crystals J Am Chem Soc 133, 12536–12543 (2011) Rasool, H I et al Continuity of graphene on polycrystalline copper Nano Lett 11, 251–256 (2011) Gao, L., Guest, J R & Guisinger, N P Epitaxial graphene on Cu(111) Nano Lett 10, 3512–3516 (2010) Kim, D W., Kim, Y H., Jeong, H S & Jung, H.-T Direct visualization of large-area graphene domains and boundaries by optical birefringency Nature Nanotechnol 7, 29–34 (2012) Acknowledgements This work was supported by the Star Faculty programme (2010-0029653), the International Research and Development programme (2011-00242) and the WCU programme (R31-2008-10029) of the NRF of Korea funded by MEST Author Contributions D.L.D and G.H.H contributed equally to this work in experiment planning, experiment measurements, data analysis and manuscript preparation S.M.L performed the theoretical calculations F.G prepared the samples for TEM measurements The copper grain size was characterized by H.K SEM and EDS were performed by E.S.K and J.W.J S.H.C and S.C.L designed the ultraviolet chamber and humidity controller S.T.K and S.C.L performed conductance AFM K.P.S designed and performed the bending test Graphene on nickel was prepared by S.J.C S.J.Y performed recovery of sheet resistance after ultraviolet exposure Q.H.T prepared the graphene samples for all the experiments The TEM images were taken by M.H.P S.M.L and J.Y.C contributed to the manuscript preparation Y.H.L contributed to experiment planning, data analysis and manuscript preparation Author Information Reprints and permissions information is available at www.nature.com/reprints The authors declare no competing financial interests Readers are welcome to comment on the online version of the paper Correspondence and requests for materials should be addressed to Y.H.L (leeyoung@skku.edu) 1 O C T O B E R 2 | VO L | N AT U R E | ©2012 Macmillan Publishers Limited All rights reserved RESEARCH LETTER METHODS Synthesis of graphene on copper Graphene was synthesized on 70-mm-thick copper foil (Wacopa) using atmospheric pressure chemical vapour deposition (APCVD) The temperature in the chamber was elevated to the growth temperature within 40 min, and the samples were annealed at the growth temperature for 30 with gas flows of 200 cubic centimetres at STP per minute (cm3 STP min21) of H2 and 1,000 cm3 STP min21 of Ar Monolayer graphene was synthesized by flowing cm3 STP min21 of H2 and 10 cm3 STP min21 of CH4 with 1,000 cm3 STP min21 of Ar for The sample was then cooled to room temperature while maintaining 1,000 cm3 STP min21 of Ar Synthesis of graphene on nickel The Ni substrate of 300 nm thickness was deposited on a SiO2/Si substrate by an electron-beam evaporator This substrate was placed in a rapid thermal chemical deposition chamber The temperature was increased from room temperature to 950 uC over with H2 gas flow To synthesize few-layer graphene, a mixed gas of C2H2/H2 (2/45) and a growth time of were used After growth, the gas supply was turned off, and the chamber was cooled down to 150 uC at a cooling rate of 160 uC min21 Around ten layers of graphene were synthesized Optical microscopy and Raman spectroscopy Optical microscopy (1003 magnification, Olympus, numerical aperture 0.9) was used to obtain images of the surface morphologies of the graphene/Cu samples Two-dimensional confocal Raman mapping (CRM 200, Witec) was also performed using a doubled Nd:YAG laser (532 nm) with mW power to confirm the optical image results The scan image was obtained at 100 100 pixels with a grating of 600 grooves per mm to yield a spectral resolution of cm21 The accumulation time for each spectrum was 0.3 s for image scanning and 30 s for a single spectrum A sum filter was used to extract the D-band (1,300–1,400 cm21), G-band (1,540–1,640 cm-1) and G9-band (2,640–2,740 cm21) distributions after copper background subtraction was performed Atomic force microscopy (AFM) The AFM images were obtained using a SPA400 system (Seiko) in tapping mode A NSC14-type silicon tip (MikroMasch) with an approximately 10-nm tip radius was used In general, the force constant and resonant frequencies of the tips were approximately N m21 and 160 kHz, respectively The conductance and morphology mappings of graphene monolayer on a Cu substrate were carried out using a Raman-AFM system from NT-MDT The system was run in contact mode for both mappings We used an Au-coated tip that had a radius of curvature of approximately 10 nm and an electrical resistivity of 0.025 V cm The force applied to the AFM tip was about 3.2 mN, which was under precise control through a feedback loop during the scanning For the current measurement, we applied V d.c to the Cu substrate and grounded the tip The scan speed was 4.3 mm s21 The morphology and electrical current were obtained simultaneously at each pixel The resolution of all the images was 512 512 d.p.i During the characterization, room temperature and humidity level were 21 uC and 50%, respectively Scanning electron microscopy (SEM) A field-emission scanning electron microscope (FESEM; JSM7000F, Jeol) was used to examine the surface morphology of the samples at different accelerating voltages to obtain a high level of contrast at different magnifications An X-Max silicon drift detector was used for energy dispersive spectroscopy mapping for a duration of one hour in a FESEMJSM7600F system An accelerating voltage of keV was used to obtain sufficiently high signals while retaining the sensitivity to the sample surface X-ray photoelectron spectroscopy (XPS) XPS was performed using an Al Ka X-ray source (XPS, ESCA2000, VG Microtech) C 1s, Cu 2p, O 1s peak data were collected to analyse the extent of oxidation of the graphene and the underlying Cu foil Transmission electron microscopy (TEM) The graphene was transferred to a carbon grid for TEM using two steps: the graphene was first transferred from the copper foil to a silicon substrate using the poly(methyl methacrylate) (PMMA)supported layer method and then transferred from the silicon substrate to a TEM grid using the direct transfer method25 PMMA liquid (MicroChem, 950 PMMA C4) was spin-coated onto graphene/Cu at 1,000 r.p.m for 60 s The copper foil was then etched away using an FeCl3 solution for 40 The PMMA/graphene was then placed onto SiO2 after rinsing with deionized water, and the PMMA was removed using acetone Isopropanol was then dropped on the graphene/SiO2 surface, and the TEM grid was attached onto the graphene/SiO2 After drying, the graphene strongly adhered to the TEM grid A diluted HF solution was then used to slowly etch the SiO2 and leave the TEM/graphene floating on the solution surface After rinsing off HF with deionized water, the graphene/TEM grid was dried for before collecting TEM images A TEM grid with a carbonsupported thin film (PELCO, 200 mesh, carbon type B) was used to collect selective area electron diffraction (SAED) patterns The use of the film-type TEM grid was necessary to reduce the possibility of breaking the monolayer graphene The SAED pattern was collected at 1.2 mm, that is, the maximum aperture size, using an HR-TEM instrument (JEM2100F, Jeol) at 200 keV Sheet resistance measurement The graphene was transferred onto a silicon wafer using a PMMA-supported layer as previously described For non-destructive tests, ultraviolet-treated graphene/Cu was transferred to a SiO2 substrate.The resistance was measured using the four-point probe method and a Keithley 2000 Multimeter Density functional theory (DFT) calculations We performed density functional theory calculations within generalized gradient approximation26 as implemented in DMol3 code27 All electron Kohn-Sham wavefunctions were expanded in a local atomic orbital basis set with each basis function defined numerically on an atomic centred spherical mesh Double numeric polarized basis sets with polarization were used for all elements We used slab geometry of 8 repeating graphene units in x- and y- directions, containing 128 carbon atoms, and applied a periodic boundary condition in three dimensions After cell relaxation, the supercell size ˚ 19.68 A ˚ and 10 A ˚ of vacuum in the z-direction was used The becomes 19.68 A ˚ 21, the maxmaximum force allowed during geometry optimization was 0.1 eV A ˚ imum displacement was 0.005 A and the total energy change was 1025 eV The damped atom-pairwise dispersion corrections of the form C6R26 were also considered for calculations28 A gamma point irreducible Monkhorst-Pack k-point grid sampling29 was used for structural relaxations The energy convergence was checked using a more refined 4 k-point sampling, and the energy difference was less than meV per atom Then a Stone–Wales (SW) defect was generated and the geometry was again optimized within the criteria mentioned above The formation energy of an SW defect was 5.13 eV For the geometries with adatoms, we added an adatom onto the graphene surface with an SW defect at an intended position and then the geometry was optimized repeatedly for every adsorption point From fully optimized geometries of reactants and products, we considered diffusion pathways of the penetration of H and O atoms or an OH molecule In order to find the diffusion barriers, transition state search routines of the linear synchronous transit (LST) method30 and the quadratic synchronous transit (QST) methods were used LST started with a single LST maximization, bracketing the maximum between the reactants and product, and was then followed by an energy minimization in directions conjugate to the reaction pathway This yielded a structure lower in energy and closer to the true transition state than a simple LST method Minimization continued until an energy minimum was reached or the number of conjugate directions was exhausted The LST approximation obtained in that way was used to perform a QST maximization The QST method interpolated the reaction pathway among three structures; an intermediate geometry was required in addition to the reactant and product structures From the QST maximized point, another conjugate gradient (CG) minimization was performed The cycle was repeated until a stationary point was located or the number of allowed QST steps was exhausted By calculations of the LST-CG then the QST-CG routine repeatedly, the transition state can be defined In order to check the Cu substrate effect, we calculated the diffusion barrier height through OH-functionalized SW defects in the presence of copper substrate This was done by placing a Cu(111) surface below the graphene layer We first optimized the Cu bulk system using a PBE functional with 12 12 12 Monkhorst-Pack k-point grid sampling, and the pseudopotential basis sets From the optimized bulk Cu, we generated two layers of Cu(111) surface with 8 surface units We put in a considerable size of vacuum in order to add graphene with admolecules for further calculations The supercell size was ˚ , the same as that of graphene in plane directions The 19.68 19.68 20.00 A Cu(111) surface was under compressive strain of 4.8%, due to the lattice mismatch ˚ between the Cu(111) surface and graphene The graphene layer was placed 3.2 A above the Cu layer The electron density for transition states obtained without a Cu layer was optimized again in the presence of a Cu surface without relaxing ions The total energy difference between the transition state and the reference state geometries has been defined as the diffusion barrier for each species diffusing 25 26 27 28 29 30 Regan, W et al A direct transfer of layer-area graphene Appl Phys Lett 96, 113102 (2010) Perdew, J P., Burke, K & Ernzerhof, M Generalized gradient approximation made simple Phys Rev Lett 77, 3865–3868 (1996) Delley, B An all electron numerical method for solving the local density functional for polyatomic molecules J Chem Phys 92, 508–518 (1990) Tkatchenko, A & Scheffler, M Accurate molecular Van Der Waals interactions from ground-state electron density and free-atom reference data Phys Rev Lett 102, 073005 (2009) Monkhorst, H J & Pack, J D Special points for Brillouin-zone integrations Phys Rev B 13, 5188–5192 (1976) Halgren, T A & Lipscomb, W N The synchronous-transit method for determining reaction pathways and locating molecular transition states Chem Phys Lett 49, 225–232 (1977) ©2012 Macmillan Publishers Limited All rights reserved ... was cooled down to 150 uC at a cooling rate of 160 uC min21 Around ten layers of graphene were synthesized Optical microscopy and Raman spectroscopy Optical microscopy (1003 magnification, Olympus,... unchanged On the other hand, the Cu grain boundary was initially grooved to a depth of 15 nm, and this depth was not altered by oxidation (Fig 3c bottom) This observation supports previous reports of... microscope (Fig 1a) In general, GGBs are not visible using optical microscopy because of the nanoscale width of the boundaries2 –8 Although copper grain boundaries were clearly observed in the optical