Tuning adlayer-substrate interactions of graphene/h-BN heterostructures on Cu(111)–Ni and Ni(111)–Cu surface alloys

The evolution of the interface and interaction of h-BN and graphene/h-BN (Gr/h-BN) on Cu(111)–Ni and Ni(111)–Cu surface alloys versus the Ni/Cu atomic percentage on the alloy surface were comparatively studied by the DFT-D2 method, including the critical long-range van der Waals forces. Our results showed that the interaction strength and interface distance of Gr/h-BN/metal can be distinctly tuned by regulating the chemical composition of the surface alloy at the interface. The initially weak interaction of h-BN/Cu(111)–Ni increased linearly with increasing Ni atomic percentage, and the interface distances decreased from ∼3.10 to ∼2.10 Å. For the h-BN/Ni(111)–Cu interface, the strong interaction of the NtopBfcc/hcp stacking decreased sharply with increasing Cu atomic percentage from 0% to 50%, and the interface distances increased from ∼2.15 to ∼3.00 Å; meanwhile, the weak interaction of the BtopNfcc/hcp stacking decreased slightly with increasing Cu atomic percentage. The absorption of graphene on h-BN/Cu(111)–Ni with BtopNhollow/NtopBfcc and BtopNhollow/BtopNfcc stacking was more energetically favorable than that with NtopBhollow/NtopBfcc and NtopBhollow/BtopNfcc at Ni atomic percentages under 75%, while the interaction energy of graphene on h-BN/Cu(111)–Ni increased sharply at Ni atomic percentages higher than 75% for the BtopNhollow/NtopBfcc and NtopBhollow/NtopBfcc stacking. In contrast, the interaction between graphene and the h-BN/Ni(111)–Cu surface increased sharply at Cu atomic percentages lower than 25% and decreased sharply at Cu atomic percentages higher than 75%. The interaction energies were higher when the percentage of Cu atom was between 25% and 75%. The analysis of charge transfer and density of states provided further details on the changing character and evolution trends of the interactions among graphene, h-BN, and Cu–Ni surface alloy versus the Ni/Cu atomic percentage.


Introduction
Graphene/hexagonal boron nitride (Gr/h-BN) heterostructures have attracted much interest due to their intriguing electronic and mechanical properties. 1,2 Great effort has been devoted to growing Gr/h-BN heterostructures with various vertically stacked or in-plane pieced patterns by chemical vapor deposition (CVD) method on various metal substrates. [2][3][4][5][6] In general, Gr/h-BN heterostructures can be grown on weakly binding metal surfaces such as Cu, 1,2,7-9 Ir, 5 Rh, 4 and Pt(111), 10 resulting in a heteroexpitaxial growth mechanism of h-BN growth along the graphene edge 1,8 or graphene nucleation at the corners of the triangular h-BN grains. 3 In this case, the common feature of these quasi-free-standing Gr/h-BN heterostructures completely remains their intrinsic electronic properties due to the weak interfacial binding by Pauli exclusion and van der Waals (vdW) attraction. 5,10 In contrast, Gr/h-BN heterostructures grown on strongly interacting metal substrates, such as Ni, 11,12 Ru, [13][14][15] and Re(111), 6 result in the coexistence of perfectly patched Gr/h-BN heterostructures linked with predominant zigzag-type boundaries 6,12 because the graphene and h-BN favor growth into separated domains. In addition, the intrinsic electronic properties of the Gr/h-BN heterostructures are almost completely inhibited due to the strong interfacial chemical bonds and charge transfer. 6 In this regard, the complex growth mechanisms and electronic properties of the Gr/h-BN heterostructures strongly depend on the interfacial interaction among graphene, h-BN and the metal substrate. Therefore, it is highly desirable to tune the interfacial interaction strength of graphene, h-BN and the metal substrate to facilitate controllable growth and electronic properties of the Gr/h-BN heterostructure.
One possibility for tuning the interfacial interaction strength of a Gr/h-BN heterostructure is changing the chemical composition of the metal surface alloy at the interface. Recently, a single-layer Gr/h-BN in-plane heterostructure was successfully synthesized on Cu-Ni alloy substrate by a two-step low pressure CVD method. 3 The Cu-Ni alloy substrate showed excellent catalytic performance, which not only enhanced the decomposition capability of polyaminoborane residues and the crystal quality of h-BN, but also eliminated random nucleation and promoted the growth of graphene through isothermal segregation. On the Cu-Ni alloy surface, graphene nucleated only at the top corners of the triangular h-BN grains and grew along the edge orientation of the as-formed h-BN with a fast growth rate. Subsequently, h-BN/graphene vertical stacked heterostructures were also successfully synthesized on the Cu-Ni alloy substrate by a CVD method in the same group. 16 The interface and interactions of Gr/h-BN on pure Cu(111) and Ni(111) substrates have been studied in recent experimental and theoretical studies, 7,11,12 which revealed the difference in the growth mechanisms and interfacial properties of the Gr/h-BN heterostructure on the weakly coupling Cu(111) and strongly interacting Ni(111) surfaces. However, few relevant experimental and theoretical investigations have been reported on the metal surface alloy, which is undoubtedly crucial to better understand the growth and electronic properties of Gr/h-BN heterostructures.
In our recent study, the interface interaction and properties of the Gr/h-BN heterostructures on pure Cu(111) and Ni(111) surfaces were investigated. 17 The results showed that h-BN and Gr/h-BN have two typical types of interactions, weakly coupled and strong, with pure metal substrates of Cu(111) and Ni(111). The N top B hcp/fcc stacking congurations of h-BN and Gr/h-BN on Ni(111) were strong chemisorption, while the B top N hcp/fcc stacking congurations on Ni(111) and both N top B hcp/fcc and B top N hcp/ fcc congurations on Cu(111) were weak physisorption. In this study, we would like to further attempt to tune the two typical interfacial interactions and surface electronic structures by regulating the chemical compositions of the surface alloys at the interface. Therefore, two types of Cu(111)-Ni and Ni(111)-Cu surface alloys with ve representative Cu/Ni ratios were chosen as representatives of surface alloys to study the evolution of the interfacial interaction and properties with various Gr/h-BN heterostructures. 3,16 The goal of this work was to elucidate the nature of the interfacial interactions and electronic properties of Gr/h-BN heterostructures on Cu-Ni surface alloys and further elaborate the inuence trends of the surface alloys on the interfacial interactions and transport properties by ne-tuning the atomic percentages of Ni and Cu on the surface of Cu(111) and Ni(111).

Computational methods and models
All spin-polarized density functional theory (DFT) calculations were performed using the Vienna Ab initio Simulation Package (VASP). 18 The projector augmented wave (PAW) pseudopotential 19,20 was used for the electron-ion interactions and Perdew-Burke-Ernzerhof (PBE) generalized gradient approximation (GGA) was used for the exchange-correlation functional. [21][22][23] Long-range dispersion corrections were considered within the DFT-D2 method. The dispersion coefficients C 6 and vdW radii R 0 for B, C, N, Ni, and Cu used in our DFT-D2 method were taken from previous work. 24,25 The scale factor S 6 was 0.75. 26,27 The energy cutoff was set to 400 eV in the plane-wave basis set, and all calculations used a convergence criterion of 10 À6 eV.
The pure metal surfaces were modeled by six-layer Cu(111) and Ni(111) periodic slab with the lowest two layers xed at their equilibrium bulk phase positions, while the upper four layers were allowed to relax. The surface alloys of Cu(111)-Ni and Ni(111)-Cu were built by substituting the Cu/Ni atoms on the topmost layer of optimized p(2 Â 2) Cu(111) and Ni(111) with the Ni/Cu atoms. In the surface layers of Cu(111)-Ni and Ni(111)-Cu, the number ratios of Ni to Cu or Cu to Ni atoms are 0/1, 1/3, 1/1, 3/1, and 1/0, corresponding to Ni/Cu concentrations of 0%, 25%, 50%, 75%, and 100%, respectively.
When adsorbing the h-BN layer on the Cu (111) Fig. 1a. The interaction energies and interlayer vertical distances are summarized in Fig. 2 and Table S1 The results in Fig. 2 and Table S1 † show that the interaction energies and interlayer distances of h-BN on the Cu(111)-Ni and Ni(111)-Cu surface alloys mainly depended on the Ni/Cu atomic percentages of the surface alloys and the stacking conguration. Fig. 2a shows that the interaction energy of monolayer h-BN increased linearly with increasing Ni atomic percentage on the Cu(111)-Ni alloy surface. The interaction energies of h-BN on the Cu(111)-Ni surface increased linearly from À0.53 to    Fig. 3a, the interaction energy of h-BN on Cu(111)-Ni(100%) is À0.80 eV per BN, which is signicantly higher than that on pure Ni(111) by $0.15 eV per BN. The corresponding interlayer distance is 2.09Å, which is shorter than that on pure Ni(111) by only 0.05Å. However, for the b N top B fcc stacking, as illustrated in Fig. 3b, the interaction energy of h-BN on Cu(111)-Ni(100%) is about À0.61/0.62 eV per BN, which is slightly lower than that on pure Ni(111) by about 0.03 eV per BN. Meanwhile, the corresponding interlayer distance is 2.71Å, which is longer than that on pure Ni(111) by 0.57/0.58Å. However, the interaction energy proles of h-BN adsorption in the B top N fcc stacking geometries exhibit only one minimum at a separation of 2.95Å for the weak physisorption, as illustrated in Fig. 3c. The corresponding adsorption energy is À0.50 eV per BN, which is higher than that on pure Ni(111) by only 0.08 eV per BN, and the corresponding interlayer distance is 2.95Å, which is shorter than that on pure Ni(111) by only 0.06Å.
On the Ni(111)-Cu surface, Fig. 2b shows that the interaction energies of h-BN on the Ni(111)-Cu surface decrease sharply from À0.65 to À0.50 eV per BN as the Cu atomic percentage increases from 0% to 50% for the N top B fcc and N top B hcp stacking with strong interaction. Accordingly, the interlayer vertical distances, d BN-M , from the h-BN layer to the topmost layer of Ni(111)-Cu increase gradually from 2.14 to 2.90Å. Subsequently, the interaction energies decrease slightly from À0.50 to À0.47 eV per BN at Cu surface concentrations higher than 50%, and the interlayer distances from the h-BN layer to the Ni(111)-Cu surface only increase by about 0.06Å. For the B top N fcc and B top N hcp stacking with weak interaction, the interaction energies of monolayer h-BN on the Ni(111)-Cu surface decrease slightly from À0.43 to À0.40 eV per BN as the Cu atom concentration increases from 0% to 100% on the alloy surface, and the interlayer vertical distances from the h-BN layer to the Ni(111)-Cu surface are only increased by about 0.06Å. Similarly, the interaction energies of the N top B fcc and N top B hcp stacking are higher than that of the B top N fcc and B top N hcp stacking when the monolayer h-BN adsorbs on the Ni(111)-Cu alloy surfaces. Additionally, the difference in the adsorption energies and interlayer distances between the FCC and HCP congurations is negligible for monolayer h-BN on the Cu(111)-Ni and Ni(111)-Cu alloy surfaces.
In addition, signicant distortion occurs for the interfacial layers when adsorbing the monolayer h-BN on Ni(111)-Cu with 25% Cu, as shown in Table S1 †   When the Cu atomic percentage is 50%, 75% or 100%, Table S1 † shows that the interfacial layers become less distorted for N top B fcc and N top B hcp stacking because the interfacial interaction decreases with increasing interlayer vertical distance. In addition, the interaction energy and interlayer distance are not signicantly different for h-BN on both the Ni(111)-Cu(100%) and pure Cu(111) surfaces. For the N top B fcc and N top B hcp stacking, the interaction energies of h-BN on Ni(111)-Cu(100%) are À0.47 eV per BN, which is lower than that on pure Cu(111) by 0.06 eV per BN, and the corresponding interlayer distances are $2.96Å, which is longer than that on pure Cu(111) by only 0.03Å. Differently, for the B top N fcc and B top N hcp stacking, the interaction energy of h-BN on Ni(111)-Cu(100%) is À0.40 eV per BN, which is lower than that on pure Cu(111) by about 0.04 eV per BN; however, the corresponding interlayer distance is $3.07 A, which is shorter than that on pure Cu(111) by only 0.04-0.01 A, respectively.
3.1.2. Interfacial bonding and charge transfer. The difference in the interaction energy and interlayer distance results in a distinct difference in the interfacial properties, as illustrated in Fig. 3 and 4, which can be further validated by analyzing the charge density distribution and density of states. For the a N top B fcc stacking with strong interaction of h-BN on the Cu(111)-Ni(100%) surface, the side view in Fig. 3d shows remarkable charge transfer in the interface between h-BN and Cu(111)-Ni(100%), resulting in a chemical bond between the N atom of h-BN and the Ni atom on the Cu(111)-Ni(100%) surface. The top view in Fig. 3d shows that the charge density increases signicantly at the B atoms of the h-BN layer, while it decreases at the N atoms. These results indicate that charge mainly transfers from Ni atoms to the B atoms through the N atoms of the h-BN layer. The Bader charge analysis further conrms that the N atoms have À2.12 e per atom, which is less than that in the isolated h-BN by only 0.01 e per atom, while the B atoms have +2.06 e per atom, which is more than that in the Similarly, for the N top B fcc stacking with strong interfacial interaction between h-BN and Ni(111)-Cu(25%), the side and top views in Fig. 4c show that remarkable charge transfer also occurs in the interface between h-BN and Ni(111)-Cu(25%), resulting in strong chemical bonds between the N atoms of h-BN and the Ni or Cu atoms on the Ni(111)-Cu(25%) surface. The charge mainly transfers from the Ni and Cu atoms to the B atoms through the N atoms of the h-BN layer; as a result, the charge density increases signicantly at the B atoms while it decreases at the N atoms of the h-BN layer, as shown in the top view of Fig. 4c. The Bader charge analysis further conrms that the B atoms have +2.10 e per atom, which is more than that in the isolated h-BN by 0.03 e per atom, and the N atoms have À2.14 e per atom, which is more than that in the isolated h-BN For h-BN on Cu(111)-Ni(100%) with strong interfacial interaction, the PDOSs of the a N top B fcc stacking in Fig. 5c show higher charge densities of states around the Fermi level in comparison to the pristine PDOS of free h-BN in Fig. S1. † Frontier orbital theory indicates that a system with higher frontier electron density is chemically reactive. In this regard, the h-BN on Cu(111)-Ni(100%) with the a N top B fcc stacking would be more chemically active than the free h-BN due to the interfacial interaction and charge transfer. In addition, several new peaks appear around 4.8, 2.0, 0.2, À0.5, and À2.0 eV for the a N top B fcc stacking on the curves of the N-p and Ni-d surface states, as illustrated in Fig. 5c. These new peaks indicate strong N-Ni chemical bonding across the interface, which is caused by the hybridization between the N-p orbitals of h-BN and the 3d orbitals of Ni in the Cu(111)-Ni(100%) surface. In contrast, for the b N top B fcc and B top N fcc stacking with medium and weak interfacial interactions, the charge densities near the Fermi level are remarkably lower than those of the a N top B fcc stacking, as shown in Fig. 5d   interfacial interactions, the PDOSs show a lower charge density of states near the Fermi level, and there is no new peak and orbital hybridization, as illustrated in Fig. 6c. Similarly, for both the N top B fcc and B top N fcc stacking of h-BN on the Ni(111)-Cu(100%) surface, there is almost zero electron density near the Fermi level and no new peak/orbital hybridization compared with the pristine PDOS of free h-BN, as shown in Fig. 6b and d, implying lower chemical reactivity and weak interactions between the p-orbitals of h-BN and the 3d orbitals of the Ni(111)-Cu(100%) surface. Overall, these results suggest that the N top B fcc stacking of h-BN on Cu(111)-Ni(100%) and Ni(111)-Cu(25%) with charge transfer and strong interfacial interaction would be more chemically reactive than the B top N fcc stacking of the h-BN on Cu(111)-Ni(100%) and both N top B fcc and B top N fcc stacking of h-BN on Ni(111)-Cu(100%). These results also imply that the chemical reactivity depends not only on the composition of the alloy surface but also on the interfacial stacking of h-BN.  Fig. 7. Fig. 7a shows that the interaction energies between the 1L-Gr/h-BN layer and Cu (111) Interestingly, for N top B hollow /N top B fcc stacking, the interaction energies between 1L-Gr/h-BN and the Cu(111)-Ni surface decrease slightly from À0.54 to À0.51 eV per supercell, followed by an increase from À0.51 to À0.54 eV per supercell, as the Ni atomic percentage increases from 0% to 75% on the Cu(111)-Ni  surface layer, as illustrated in Fig. 7a; these energies are lower than those of monolayer h-BN on the corresponding Cu(111)-Ni surfaces by about 0.05 eV per supercell. The corresponding interlayer distances of d BN-M decrease linearly from 2.90 to 2.76 A as the Ni atomic percentage increases from 0% to 75% for the N top B hollow /N top B fcc stacking; these distances are shorter than those of monolayer h-BN on Cu(111)-Ni substrate by about 0.03 A. Similarly, when the Ni atomic percentage is higher than 75%, the interaction energies of 1L-Gr/h-BN with the Cu(111)-Ni surface increase sharply from À0.54 to À0.82 eV per supercell for the N top B hollow /N top B fcc stacking, and the interlayer distances from the h-BN layer to the Cu(111)-Ni surface decrease sharply from 2.76 to 2.07Å.

Graphene on the h-BN/Cu(111)-Ni and h-BN/Ni(111)-Cu
The same change trend in the interaction energy was found for B top N hollow /B top N fcc and N top B hollow /B top N fcc stacking. As the Ni atomic percentage increases from 0% to 100% in the Cu(111)-Ni surface, the interaction energies between the 1L-Gr/ h-BN layer and Cu(111)-Ni surface increase linearly from À0.46 to À0.51 eV per supercell for B top N hollow /B top N fcc stacking, as illustrated in Fig. 7a, which are similar to those of monolayer h-BN on the corresponding Cu(111)-Ni surfaces. The energies also increase linearly from À0.39 to À0.45 eV per supercell for N top B hollow /B top N fcc stacking, as illustrated in Fig. 7a, which are lower than those of B top N hollow /B top N fcc stacking by about 0.06 eV per supercell. The interlayer distances from the h-BN layer to the Cu(111)-Ni surface, d BN-M , decrease linearly from 3.04 to 2.90Å with increasing Ni concentration from 0% to 100% for N top B hollow /B top N fcc and B top N hollow /B top N fcc stacking, which are shorter than those of monolayer h-BN on the Cu(111)-Ni surface by a range of 0.02-0.06Å.
In contrast, Fig. 7b shows that with increasing Cu atomic percentage in the Ni(111)-Cu alloy surface, the interaction energies between the 1L-Gr/h-BN layer and Ni(111)-Cu surface decrease signicantly for both B top N hollow /N top B fcc and N top -B hollow /N top B fcc stacking, while they decrease slightly for both B top N hollow /B top N fcc and N top B hollow /B top N fcc stacking. In comparison to monolayer h-BN on the Ni(111)-Cu alloy surface, the interaction energies between the 1L-Gr/h-BN layer and Ni(111)-Cu surface are always higher than those of the corresponding monolayer h-BN with the Ni(111)-Cu surface due to the adsorption of the upper graphene layer.
On the strong interaction Ni(111) surfaces, the B top N hollow / N top B fcc and N top B hollow /N top B fcc stacking of 1L-Gr/h-BN have almost identical interaction energies of À0.79 and À0.80 eV per supercell and interlayer distances of 2.13 and 2.12Å, respectively. As the Cu atomic percentage increases from 0% to 100% on the Ni(111)-Cu surface layer, the interaction energies between 1L-Gr/h-BN and the Ni(111)-Cu surface decrease signicantly from À0.79 to À0.57 eV per supercell for the B top -N hollow /N top B fcc stacking and from À0.80 to À0.52 eV per supercell for the N top B hollow /N top B fcc stacking, as illustrated in Fig. 7b, which are higher than those of the monolayer h-BN on the corresponding Ni(111)-Cu surface (about 0.14 to 0.10 and 0.15 to 0.05 eV per supercell, respectively). The interaction energies between 1L-Gr/h-BN and the Ni(111)-Cu surface decrease slightly from À0.52 to À0.49 eV per supercell for B top N hollow /B top N fcc stacking, as illustrated in Fig. 7b, higher than those of h-BN on the Ni(111)-Cu surface by about 0.10 eV per supercell; meanwhile, they also decrease slightly from À0.46 to À0.42 eV per supercell for N top B hollow /B top N fcc stacking, as illustrated in Fig. 7b, which are also higher than those of monolayer h-BN on the corresponding Ni(111)-Cu surface by about 0.03 eV per supercell but lower than those of the B top -N hollow /B top N fcc stacking by about 0.07 eV per supercell.
Interestingly, when graphene adsorbs on h-BN/Ni (111)-Cu(50%), the interlayer distances between the h-BN layer and the Ni(111)-Cu(50%) surface are reduced from 2.88 Cu /2.90 Ni to 2.51 Cu /2.24 NiÅ for B top N hollow /N top B fcc stacking and from 2.88 Cu /2.90 Ni to 2.55 Cu /2.25 NiÅ for N top B hollow /N top B fcc stacking, as illustrated in Fig. 8, Tables S1 and S2. † The corresponding interaction energies between 1L-Gr/h-BN and the Ni(111)-Cu(50%) surface increase from À0.50 to À0.64 eV per supercell for B top N hollow /N top B fcc stacking and from À0. When the Ni atomic percentage is less than 75% on the Cu(111)-Ni alloy surface, the interaction energy between the graphene layer and h-BN/Cu(111)-Ni surface only depends on the stacking conguration of graphene and the h-BN layers. Fig. 9a shows that the interaction energies between graphene and the h-BN/Cu(111)-Ni surface range from À0. 28  When the Ni atomic percentage is higher than 75% on the Cu(111)-Ni alloy surface, the interaction energy between graphene and the h-BN/Cu(111)-Ni surface depends not only on the stacking conguration of the graphene and h-BN, but also on the stacking conguration of h-BN and Cu(111)-Ni. Fig. 9a shows that the interaction energy between graphene and h-BN/ Cu(111)-Ni increases sharply from À0. 29  On the Ni(111)-Cu alloy surface, the interaction energies between graphene and the h-BN/Ni(111)-Cu surface are distinctly higher than those of the corresponding freestanding Gr/h-BN. Interestingly, when the Cu atomic percentage is less than 25% on the Ni(111)-Cu alloy surface, Fig. 9b shows that the interaction energies between graphene and the h-BN/Ni(111)-Cu surface increase sharply from À0.31 to À0.61 eV per supercell for N top B hollow /B top N fcc stacking, from À0.36 to À0.64 eV per supercell for N top B hollow /N top B fcc stacking, and from À0.36 to À0.67 eV per supercell for B top N hollow /B top N fcc stacking, while it increases slightly from À0.63 to À0.70 eV per supercell for B top N hollow /N top B fcc stacking.
With increasing Cu atomic percentage from 25% to 75% on the Ni(111)-Cu alloy surface, Fig. 9b shows that the interaction energies between the graphene layer and h-BN/Cu (111)

Discussion
Recent experimental and theoretical studies have indicated that the interfacial interaction strength among graphene, h-BN, and the metal substrate plays important roles in the controllable growth and electronic properties of Gr/h-BN heterostructures. [4][5][6] Experimental studies have reported that Gr/h-BN growth on strongly interacting metal substrates results in the coexistence of perfectly patched Gr/h-BN heterostructures linked with predominant zigzag-type boundaries, 6,12 while the Gr/h-BN growth on the weakly binding metal surfaces leads to a hetero-expitaxial growth mechanism of h-BN growth along the graphene edge 1 or graphene nucleation at the corners of the triangular h-BN grains. 3 It is important to be able to tune the interaction among graphene, h-BN, and the metal surface at different growth stages to achieve controlled growth, which can be achieved through doping metal surfaces to form metal alloys.
The intention of this work was to nd a general guideline for the selection and tuning of metal surface alloys for controlled growth of Gr/h-BN and also to provide some understanding of the graphene-h-BN-metal interface for potential electronic device design. Thus, we further extended our study to investigate the interfacial structures and interactions of Gr/h-BN layers with the bimetallic Ni/Cu(111) surface alloys, and we focused on the overall interaction among large graphene, h-BN layer and metal surface alloys in addition to the individual steps at different Gr/h-BN growth stages. Our results from the interaction energy, geometric structure, charge transfer, and DOS analyses demonstrate that the interaction strength among graphene, h-BN, and the bimetallic alloy surface can be tuned selectively by reasonably regulating the atomic percentage on the alloy surface. The initially weak interfacial interaction of h-BN/Cu(111) can be enhanced substantially by introducing a Ni surface. In contrast, the initially strong interfacial interaction of h-BN/Ni(111) can be reduced successfully by introducing a Cu surface.
Recently, Lu et al. 3 demonstrated successful CVD synthesis of controlled growth of high-quality Gr/h-BN in-plane and stacked heterostructures on Cu-Ni alloy. They veried that the introduction of nickel to a copper substrate not only enhances the catalytic capability of decomposing polyaminoborane residues but also promotes graphene growth via isothermal segregation. Based on the ndings from this study, we proposed a new strategy to control growth of high-quality Gr/h-BN, that is, to design surface alloy catalysts by synergetic combination of the distinct catalytic capabilities of Cu and Ni and the wellknown segregation phenomenon in Cu/Ni binary alloy. The Ni (Cu) atoms introduced to the surface or subsurface layers of Cu (Ni) substrates can be used as high catalytic activity sites, while the Cu atoms located at the surface or subsurface cam be employed as a favorable segregation medium, such as Cu(111)-Ni-Cu (Ni(111)-Ni-Cu). To facilitate controllable growth of Ge/ h-BN, this can be achieved through regulating the atomic percentage and thickness of the Ni-Cu surface layers in the Cu(111)-Ni-Cu or Ni(111)-Ni-Cu surface alloys; therefore, the catalytic capability and isothermal segregation of boron-nitride and carbon species on the Cu(111)-Ni-Cu or Ni(111)-Ni-Cu surface layers cam be changed to provide the desired growth conditions.

Conclusions
The evolution of the interface and interaction of the monolayer h-BN and 1L-Gr/h-BN heterostructures on Cu(111)-Ni and Ni(111)-Cu surface alloys versus Ni/Cu atomic percentage were comparatively studied by the DFT-D2 method. For monolayer h-BN on both Cu(111)-Ni and Ni(111)-Cu, the interaction energies of the N top B fcc/hcp stacking are higher than those of the B top N fcc/hcp stacking. The interaction energy of h-BN on the Cu(111)-Ni surface increased linearly with increasing Ni atomic percentage for these four stacking congurations. In contrast, the interaction energies of h-BN on Ni(111)-Cu decreased slightly as the Cu atomic percentage increased, except that the interaction energies decreased sharply as the Cu atomic percentage increased from 0% to 50% for the N top B fcc/hcp stacking. The interlayer distances between h-BN and Cu(111)-Ni decreased gradually with increasing Ni atomic percentage and increased gradually with increasing Cu atomic percentage for h-BN and Ni(111)-Cu.
The interaction strength of the 1L-Gr/h-BN heterostructure on Cu(111)-Ni and Ni(111)-Cu followed the order B top N hollow / N top B fcc > N top B hollow /N top B fcc > B top N hollow /B top N fcc > N top B hollow / B top N fcc . For both B top N hollow /N top B fcc and N top B hollow /N top B fcc stacking, the interaction energies of 1L-Gr/h-BN on Cu(111)-Ni increased sharply when the Ni atomic percentage was higher than 75%, while the interaction energies of 1L-Gr/h-BN on Ni(111)-Cu decreased signicantly when the Cu atomic percentage was higher than 50%. However, for the N top B hollow / B top N fcc and B top N hollow /B top N fcc stacking, the interaction energies changed only slightly for 1L-Gr/h-BN on Cu(111)-Ni and Ni(111)-Cu. The interaction energy of graphene on h-BN/ Cu(111)-Ni with B top N hollow stacking were higher than that with N top B hollow stacking at Ni atomic percentages under 75%, while the interaction energy of graphene on h-BN/Cu(111)-Ni increased sharply at Ni atomic percentages higher than 75% for the N top B hollow /N top B fcc and B top N hollow /N top B fcc stacking. Differently, the interaction energies between graphene and the h-BN/Ni(111)-Cu surface increased sharply at Cu atomic percentages lower than 25%, while they decreased sharply at Cu atomic percentages higher than 75%. The interaction energies were higher when the percentage of Cu atom was between 25% and 75%. These results suggest that the interfacial interactions and the properties of graphene, h-BN and Cu-Ni alloy can be regulated by tuning the Cu/Ni atomic percentage on the Cu-Ni surface alloys, which provides insight into the epitaxial growth of Gr/h-BN heterostructures and the design of Gr/h-BN-based electronic devices.

Conflicts of interest
The authors declare that there are no conicts of interest.