Subscriber access provided by Kaohsiung Medical University
Crystalline Bilayer Graphene with Preferential Stacking from Ni-Cu Gradient Alloy Zhaoli Gao, Qicheng Zhang, Carl H. Naylor, Youngkuk Kim, Irfan Haider Abidi, Jinglei Ping, Pedro Ducos, Jonathan Zauberman, Meng-Qiang Zhao, Andrew M. Rappe, Zhengtang Luo, Li Ren, and Alan T. Charlie Johnson ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.7b06992 • Publication Date (Web): 06 Mar 2018 Downloaded from http://pubs.acs.org on March 7, 2018
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
Crystalline Bilayer Graphene with Preferential Stacking from Ni-Cu Gradient Alloy Zhaoli Gao†, Qicheng Zhang†,‡, Carl H. Naylor†, Youngkuk Kim#,§, Irfan Haider Abidi‡, Jinglei Ping†, Pedro Ducos†,∥, Jonathan Zauberman†, Meng-Qiang Zhao†, Andrew M. Rappe#, Zhengtang Luo‡, Li Ren⊥, *, Alan T. Charlie Johnson†,* †
Department of Physics and Astronomy, University of Pennsylvania, Philadelphia 19104, USA
‡
Department of Chemical and Biomolecular Engineering, Hong Kong University of Science and
Technology, Clear Water Bay, Kowloon, Hong Kong #
The Makineni Theoretical Laboratories, Department of Chemistry, University of Pennsylvania,
Philadelphia, Pennsylvania 19104-632, USA §
Department of Physics, Sungkyunkwan University, Suwon 16419, Korea
∥
Departamento de Física, Universidad San Francisco de Quito, Quito 170901, Ecuador
⊥
School of Materials Science and Engineering, South China University of Technology,
Guangzhou 510006, PR China KEYWORDS: bilayer graphene, single-crystal, Ni-Cu gradient alloy, Bernal and 30 ° stacking order, high-yield synthesis
ABSTRACT 1 ACS Paragon Plus Environment
ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 24
We developed high-yield synthesis of highly crystalline bilayer graphene (BLG) with two preferential stacking modes using a Ni-Cu gradient alloy growth substrate. Previously reported approaches for BLG growth include flat growth substrates of Cu or Ni-Cu uniform alloys and “copper pocket” structures. Use of flat substrates has the advantage of being scalable, but the growth mechanism is either “surface limited” (for Cu) or carbon precipitation (for uniform NiCu), which results in multi-crystalline BLG grains. For copper pockets, growth proceeds through a carbon back-diffusion mechanism, which leads to the formation of highly crystalline BLG but scaling of the copper pocket structure is expected to be difficult. Here we demonstrate a Ni-Cu gradient alloy that combines the advantages of these earlier methods: the substrate is flat, so easy to scale, while growth proceeds by a carbon back-diffusion mechanism leading to high-yield growth of BLG with high crystallinity. The BLG layer stacking was almost exclusively Bernal or twisted with an angle of 30 °, consistent with first principles calculations we conducted. Furthermore, we demonstrated scalable production of transistor arrays based crystalline Bernalstacked BLG with a bandgap that was tunable at room temperature.
Bilayer graphene (BLG) consists of two graphene monolayers and is typically found in the AB form (commonly referred to as Bernal-stacked), the AA form (hexagonal on-top stacked), or in a twisted configuration, where the layers are rotated with respect to each other at a given angle (tBLG).1 BLG offers a rich array of physical properties such as a tunable band gap,2, topological valley transport,4,
5
3
van Hove singularities,6-9 coherent commensurate electronic
states,10, 11 superlattice-induced insulating states,12 and helical edge states.13 Production of BLG crystals based on mechanical exfoliation2, 3 has low yield. Preparation of BLG by stacking one graphene monolayer on top of another requires a multiple-step transfer process to yield tBLG 2 ACS Paragon Plus Environment
Page 3 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
with a controlled twist angle.12, 14, 15 Scalable synthesis of highly crystalline BLG with controlled stacking structure is thus of great significance for both basic scientific research and potential applications. Chemical vapor deposition (CVD) is a promising approach to grow large-area BLG on metal substrates (e.g., Cu, Ni/Cu alloys), and progress has been made towards growth of BLG in the Bernal stacked form.16-22 However, synthesis of BLG with large crystalline domains remains a significant experimental challenge. BLG synthesis on flat Cu and uniform Ni-Cu substrates has been reported where growth was either surface-mediated17,
22, 23
or by a carbon precipitation
mechanism (for Ni-Cu).16, 18, 21 Typically, the BLG domain size was small and variable, leading to undesirable stacking structure transitions within a BLG region.17,
18, 24-26
A recent report
detailed synthesis of single-crystal BLG with large domains through a carbon back-diffusion mechanism,27 but the growth was based on a manually constructed Cu pocket that is less suitable for industrial scale-up compared to conventional flat growth substrates. It is also highly desirable to obtain tBLG with reproducible twist angle,27-31 which is crucial for investigations of superlattice-induced transport11, 12 and helical edge states.13 Progress in the field has been slowed due to the lack of control over the grain size and interlayer stacking in CVD-grown BLG; for example, we are not aware of any reports to date of scalable fabrication of transistor arrays based on CVD grown crystalline Bernal-stacked BLG where the band gap could be measurably tuned with an applied electric field at room temperature. Here we report the use of an optimized Ni-Cu gradient alloy for high-yield synthesis of large crystalline BLG domains (200 ± 50 µm for the top graphene layer and 20 ± 4 µm for the bottom layer) with layer stacking that is almost exclusively Bernal or tBLG with a twist angle of 30 °. By proper design of the Ni-Cu gradient alloy, we enabled BLG growth through a carbon back3 ACS Paragon Plus Environment
ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
diffusion mechanism previously obtained only using a copper pocket structure.27 A scalable photolithography process was used to fabricate arrays of dual-gate field-effect transistors based on Bernal-stacked BLG. The BLG band gap was tunable at room temperature, with on/off ratio of 5.0 ± 1.1. We also identified strain solitons32 in crystalline Bernal-stacked BLG, which is useful to the study of topologically protected states and valley physics in graphene.4, 33 RESULTS Figure 1 is a schematic of the proposed growth mechanism for BLG on the Ni-Cu gradient alloy catalytic substrate. Since Ni is a more active catalyst for CH4 decomposition than Cu34 and has higher affinity for carbon,35 we expected a higher graphene nucleation and growth rate on the Nirich side of the growth substrate.36-38 The top layer of the BLG forms first over most of the surface of the Ni-rich side, while the bottom layer of BLG grows beneath it by carbon backdiffusion through the Ni-Cu foil from the Ni-poor side.26, 27 Catalytic substrates were fabricated by depositing a Ni film of known thickness on one side of a flat 25 µm thick Cu sheet. Synthesis was carried out by atmospheric pressure chemical vapor deposition (APCVD). The Ni-Cu foils were annealed at 1050 °C for 5 min to allow for Ni diffusion to create a gradient alloy substrate with a Ni-rich side and a Ni-poor side, and then 1.8 sccm methane (CH4) was introduced as the carbon source.
4 ACS Paragon Plus Environment
Page 4 of 24
Page 5 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
Figure 1. Schematic of the proposed bilayer graphene (BLG) growth on a Ni-Cu gradient alloy substrate by carbon back-diffusion. To visualize the results, as-grown samples were transferred onto 250 nm SiO2/Si substrates. For an optimized Ni-Cu gradient alloy (160 nm Ni on Cu; see below for further discussion), the Nirich side was covered by BLG (~ 20%) and monolayer graphene (MLG; ~ 80%) with a MLG nucleation density of 23 ± 7 mm-2 (Figure 2a, left panel). The Ni-poor side showed a sparse growth of discrete MLG grains with a nucleation density of 1.3 ± 0.4 mm-2 (Figure 2b, left panel). This asymmetric growth allowed for carbon back-diffusion, leading to the formation of BLG with well controlled layer number and compact hexagonal shapes. (Figure 2a, right panel). The grain size for the top (bottom) layer of BLG was 200 ± 50 µm (20 ± 4 µm). The slower growth rate of the bottom layer compared to the top layer can be attributed to the additional decomposition and diffusion energy required for the carbon source to diffuse from the Ni-poor side to the Ni-rich.27 In contrast, for growth on Cu foil with no Ni (Figure S1), graphene coverage was low (< 10% on both sides), with occasional multilayer patches but almost no BLG. 5 ACS Paragon Plus Environment
ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Figure 2. Results of BLG growth on both sides of a Ni-Cu gradient alloy substrate, after transfer onto Si/SiO2. Optical micrographs of growth on the (a) top (Ni-rich) and (b) bottom (Ni-poor) surfaces of an optimized Ni-Cu gradient alloy substrate. The Ni-rich side is almost fully covered by graphene (domain size of 200 ± 50 µm), which forms the top layer of BLG. The bottom BLG layer forms by back diffusion of carbon atoms from the Ni-poor side. The right panel of (a) shows single crystal BLG regions, where the size of the bottom layer is 20 ± 4 µm. The Ni-poor side of the substrate has a low density of monolayer graphene domains (surface coverage of ~ 18%).
6 ACS Paragon Plus Environment
Page 6 of 24
Page 7 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
The proposed growth mechanism requires different Ni concentrations at the two surfaces of the gradient alloy in order to allow for asymmetric growth of graphene and carbon back-diffusion. We tested this assumption using time-of-flight secondary ion mass spectrometry (ToF-SIMS) to measure Ni concentration as a function of depth into an optimized substrate (160 nm Ni; 5 min annealing time to create the Ni gradient). The Ni concentration was ~ 20× higher on the Ni-rich side (Figure S2), and this surface had higher graphene coverage (~100% compared to ~ 18% on the Ni-poor side), high BLG nucleation density (3800 ± 500 mm-2) and high-yield growth of crystalline BLG. Figure S3 shows that a growth time of 1.5 hour resulted in BLG coverage of over 70% with large domains ~ 25 µm in size , in sharp contrast to the growth on Cu foil without Ni where BLG was rarely found (Figure S1). Although our coverage is somewhat less than that reported earlier for BLG growth on uniform epitaxial Ni-Cu alloy on sapphire,18 our approach offers the advantage of highly-crystalline BLG with well-defined domain sizes using a substrate material of much lower cost. We also conducted similar measurements on non-optimized substrates with thinner Ni (80 nm Ni, 5 min anneal) and longer annealing time (160 nm Ni, 30 min annealing). In each case the Ni gradient was less pronounced compared to the optimized substrate, leading to the lack of asymmetric growth and lower quality BLG formation as shown in Figure S4. To further explore the role of back diffusion in this BLG growth process, we tested the hypothesis that if carbon diffusion was prevented in a region on the Ni-poor side, then on the opposite side of the substrate, BLG formation would be frustrated. To this end, we transferred a piece of MLG (5 mm × 5 mm) onto the Ni-poor side of an optimized substrate to block carbon adsorption and back-diffusion in this region, and then conducted CVD growth for an extended time of 3 hours (Figure 3). We found that the entire Ni-rich surface had a high density of 7 ACS Paragon Plus Environment
ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
multilayer graphene except for the region that was directly opposite the transferred MLG. On the Ni-rich side, this area, where carbon back diffusion was effectively blocked, was covered by MLG, with no BLG or multilayer graphene formation (Figure 3c). This demonstrates that carbon back-diffusion is required for BLG formation in our process but it remains unclear what fraction of the carbon atoms in the bottom BLG layer derive from back-diffusion rather than excess carbon absorbed on the Ni-rich side. We believe that back-diffused carbon is the majority component because if excess carbon were absorbed on the Ni-rich side, it is energetically more favorable for it to form a bottom BLG layer compared to diffusion towards the Ni-poor side.39 The complete lack of BLG formation in the control experiment suggests that back-diffusion provides the primary carbon source for the bottom layer of BLG on Ni-Cu gradient alloy.27 To further investigate the BLG growth mechanism, we conducted an experiment with a short 10minute growth. For this sample, we found nucleation of MLG (which would have become the top layer of BLG in a longer growth), and no nucleation of the bottom layer of BLG was observed (Figure S5). These results suggest that the top layer of BLG forms first, and the bottom BLG layer nucleates independently and grows beneath it by carbon back-diffusion.
Figure 3. Control experiment providing strong evidence for the back-diffusion growth mechanism. Monolayer graphene (MLG) was transferred onto the Ni-poor side of an optimized Ni-Cu gradient alloy substrate prior to CVD growth for a time of 3 hours. (a) Photograph of the 8 ACS Paragon Plus Environment
Page 8 of 24
Page 9 of 24 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
ACS Nano
Ni-poor side after CVD growth. The transferred MLG (indicated by the black arrow) has remained as a continuous film. The black box outlines a region that was partly covered by transferred MLG and partly exposed. (b) Photograph of the Ni-rich side of the substrate after CVD growth. The surface is fully covered by graphene. The black box outlines the region that is opposite the black box in panel (a). (c) Optical micrograph of the region in the black box in panel (b) after transfer onto Si/SiO2. The area opposite to the transferred MLG is covered with MLG but shows no sign of multilayer graphene formation. In contrast, nearby areas that were not opposite to the transferred MLG show abundant formation of multi-layer graphene. BLG synthesized by this approach was highly crystalline and formed preferentially as Bernal stacked BLG or tBLG with a twist angle of 30 ˚, as confirmed by optical microscopy, Raman spectroscopy and selected area electron diffraction (SAED). Figure 4a is an optical image of a crystalline domain of MLG containing multiple hexagonal BLG regions. The MLG orientation is inferred from the hexagonal edge marked with a white dashed line in the upper right corner of the image. The hexagonal bottom layer grains took on two orientations corresponding to Bernal stacked BLG (indicated by white dashed lines), and tBLG with a twist angle of 30˚ (yellow dashed lines). This identification was consistent with point Raman spectra taken in the various regions (see Figure S6) and Raman mapping of the width of the 2D band, as seen in Figure 4bc. The 2D FWHM map shows only monolayer graphene and Bernal-stacked BLG with no other apparent stacking variations, consistent with the assignment that BLG is highly crystalline. Analysis of optical micrographs showed that number of Bernal-stacked BLG regions was about the same as tBLG 30˚, which was confirmed by inspection of the 2D band in the Raman spectra8, 28, 40
taken from 60 randomly selected BLG grains. As shown in Figure 4d, 48% of the BLG
grains were Bernal-stacked (2D FWHM: ~50 cm-1, normalized 2D intensity: ~1.2), 44% were 9 ACS Paragon Plus Environment
ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
30˚ tBLG (2D FWHM: ~25 cm-1, normalized 2D intensity: ~2.4), and the remaining 8% were tBLG with angles very different from 30 ° (2D FWHM: ~35 cm-1, normalized 2D intensity:
