PCCP

View Article Online

Published on 04 September 2014. Downloaded by University of California - Riverside on 23/04/2015 05:43:09.

PAPER

Cite this: Phys. Chem. Chem. Phys.,

2014, 16, 22861

View Journal | View Issue

Tunable electronic properties induced by a

defect-substrate in graphene/BC3 heterobilayers
Sheng-shi Li, Chang-wen Zhang,* Wei-xiao Ji, Feng Li and Pei-ji Wang
We perform first-principles calculations to study the geometric, energetics and electronic properties of
graphene supported on BC3 monolayer. The results show that overall graphene interacts weakly with
BC3 monolayer via van der Waals interaction. The energy gap of graphene can be up to B0.162 eV in
graphene/BC3 heterobilayers (G/BC3 HBLs), which is large enough for the gap opening at room

Received 22nd July 2014,

Accepted 29th August 2014

temperature. We also find that the interlayer spacing and in-plane strain can tune the band gap of

DOI: 10.1039/c4cp03248a

dispersion relationship of graphene can be preserved, accompanied by a small electron eective mass,

G/BC3 HBLs eectively. Interestingly, the characteristics of a Dirac cone with a nearly linear band
and thus the higher carrier mobility is still expected. These findings provide a possible way to design

www.rsc.org/pccp

eective FETs out of graphene on a BC3 substrate.

I. Introduction
Graphene, due to its intriguing electronic properties as well as
its high carrier mobility, has attracted enormous attention
since its experimental realization in 2004.1,2 Unfortunately,
the zero-gap property of graphene makes it very dicult to
apply in electronic devices such as high-performance field
eect transistors (FETs) operating at room temperature. Considerable attempts have been made to open the band gap,38
among which the bilayer nanostructures are of particular
interest, including SiO2,9,10 SiC,11,12 MoSe2,13 WS2,14 and BN15
substrates. Despite these achievements, the search for ideal
substrates is still underway.
Recently, BC3, a two-dimensional (2D) defect-like graphene,
has been realized in an epitaxial way on the NbB2 (0001)
surface.16 The 2D BC3 is a semiconductor with an intermediate
band-gap of 0.50.66 eV predicted within local density approximation (LDA) using different basis sets.17 Chen et al.18 further
investigated the Fermi surface nesting-derived magnetic phases
in BC3, suggesting BC3 as an experimentally feasible nanosystem for exploring the nesting-driven quantum phases.
In the present work, we design new 2D G/BC3 HBLs
composed of graphene and BC3 monolayer, due to their similar
crystal geometries. It is revealed that overall graphene interacts
weakly with BC3 monolayer via van der Waals interaction. Thus,
the intrinsic properties of graphene, especially, high carrier
mobility, can be preserved in G/BC3 HBLs. We also find that the
sizeable band gap would be induced by interlayer interactions,
and all band gaps can be tuned eectively by changing the
School of Physics and Technology, University of Jinan, Jinan, Shandong, 250022,
Peoples Republic of China. E-mail: zhchwsd@163.com

This journal is the Owner Societies 2014

interlayer spacing and in-plane strain. Our work is expected to

provide an eective way to tune the electronic properties of
layered graphene-related nanoelectronic devices.

II. Computational details

All the calculations were performed by means of first-principles
calculations as implemented in the Vienna Ab Initio Simulation
Package (VASP).1921 Generalized gradient approximation (GGA)
using the PerdewBurkeErnzerhof (PBE)22,23 exchange correlation
functional was adopted to describe the exchangecorrelation
interaction, which is developed for the calculations of surface
systems. To properly take into account the van der Waals (vdW)
interactions in the structures, the DFT-D2 method24 was used
throughout all the calculations. The projector augmented wave
method25 was used to describe the electronion interaction.
The dipole corrections were also included considering the
possible charge redistribution in the HBLs. The plane-wave
cutoff energy was set to 400 eV, and a supercell with the 20
vacuum layer was used to simulate the isolated sheet. The
MonkhorstPack (MP)26 scheme was used to sample the Brillouin
zone, which adopted the 11  11  1 and 27  27  1 k-mesh for
the relaxation and static calculations, respectively. All the lattice
constants and atom coordinates were optimized until the convergence of the force on each atom is less than 0.01 eV 1.

III. Results and discussion

In order to explore the interaction between graphene and the
BC3 substrate, we consider three representational stacking
patterns as shown in the insets of Fig. 1(a)(c). P-I exhibits

Phys. Chem. Chem. Phys., 2014, 16, 22861--22866 | 22861

View Article Online

Published on 04 September 2014. Downloaded by University of California - Riverside on 23/04/2015 05:43:09.

Paper

PCCP

Fig. 1 The variation in the binding energy for the three configurations as a
function of interlayer spacing. Top views of the relaxed atomic structures
of three stacking patterns of G/BC3: (a) P-I, (b) P-II, and (c) P-III. The side
view of the relaxed atomic structure of P-II (d).

hexagonal stacking arrangement (AA) with all C atoms in

graphene right above C and B atoms in the BC3 substrate;
P-II has the Bernal stacking (AB) features, with the C atoms of
graphene right above the CB or CC hexagonal center within
BC3 monolayer, while P-III can be achieved by translating
p
graphene a distance of 3 6a, where a is the lattice constant
of graphene, with respect to BC3 along CB bond orientation
from pattern AA. After structural relaxations, we find that
graphene retains its original planar and hexagonal atomic
network, while the lattice constant of G/BC3 HBLs increases
from 2.44 to 2.565 , irrespective of the stacking types. To
evaluate the structural stability of the three stacking patterns,
the binding energies are calculated as
Eb = E(G/BC3)  E(G)  E(BC3)

(1)

where E(G/BC3), E(G) and E(BC3) are energies of G/BC3 HBLs,

isolated graphene and BC3 monolayer, respectively. The evolution of the binding energy (Eb) per C atom as a function of the
interlayer spacing is plotted in Fig. 1. We find that patterns IIII
exhibit a similar equilibrium interlayer spacing of about 3.3 ,
as shown in the side view of P-II [see the inset in Fig. 1(d)]. The
stability of the HBLs follows the order of P-II 4 P-I 4 P-III,
suggesting that the stacking pattern P-II is the most stable one.
Furthermore, the lower Eb of 0.099 eV also indicates that there
is a weak van der Waals interaction between graphene and BC3
monolayer.
Notice that the graphene retains its planar structure and the
interlayer distance is much larger than the sum of covalent
radii of C and B atoms, which means that the C atoms of
graphene and C or B atoms of BC3 are beyond the bonding
range. To explore whether the electronic structures of graphene
can be aected by the BC3 substrate, taking the energetically
favorable P-II as an example, we examine the projected band
structures of the G/BC3 HBLs, which show that the contributions from graphene and the BC3 substrate are separated, as
shown in Fig. 2(a). For comparison, we also present the band
structure of free-standing graphene (FSG) and BC3 monolayer
in Fig. 2(b) and (c). It can be seen that a metalsemiconductor
transition occurs in graphene due to the effect of the BC3
substrate. The valence band maximum (VBM) and the conduction band minimum (CBM) in G/BC3 HBLs have the typical
p and p* characters, similar to the FSG one. According to the
p-electron tight-binding (TB) model of bipartite lattices, the band
dispersion relationship near the Fermi level can be expressed as
q
jEkj  D2 
(2)
hn F k2
Here, k is the wave vector relative to the K point, vF is the
Fermi velocity, and D is the onsite energy dierence between
two sublattices. For FSG, the onsite energies of A and B

Fig. 2 (a) The band structure of P-II: the cyan, green and red represent the component of graphene, B and C atoms in the BC3 substrate, respectively.
(b) The band structure of FSG, and (c) the band structure of isolated BC3.

22862 | Phys. Chem. Chem. Phys., 2014, 16, 22861--22866

This journal is the Owner Societies 2014

View Article Online

Published on 04 September 2014. Downloaded by University of California - Riverside on 23/04/2015 05:43:09.

PCCP

Paper

sublattices are identical (D = 0), and thus results in the Diraclike linear dispersion relationship E(k) = hnF|k|. The positive
and negative solutions, which correspond to conduction and
valence bands, respectively, meet at k = 0, implying the absence
of a band gap. As shown in Fig. 2(a), when graphene is
deposited on the BC3 substrate, the symmetry of two C sublattices within graphene is broken, which induces a small gap
of 0.162 eV at the K point, which is large enough for the gap
opening at room temperature. Fig. 3 further shows the total and
partial densities of states (DOSs) for graphene and BC3 monolayer. It can be seen that the C-2p orbital within graphene
dominates the states near the Fermi level, and there is a gap
within the range of [0.08, 0.082] eV [Fig. 3(a) and (b)].
Furthermore, the bands that originated from the C-2p orbital
are quite analogous to the FSG one, except for the downshift of
top valence bands at the K point. Similarly, for the bands
contributed by BC3 monolayer, while VBM at the K point moved
up towards the Fermi level, the CBM from the B-2p orbital is
down-shifted. Thus, the gap of BC3 monolayer decreases from
the pristine value of 0.56 eV to 0.45 eV [Fig. 3(c) and (d)].
Overall, around the Fermi level, the band structures of G/BC3
HBL have the characteristic graphene feature of linear dispersions, i.e., the merits of the Dirac system, such as the high
Fermi velocity and large carrier mobility, may be well retained
in the G/BC3 HBLs.
As mentioned above, the band gap is attributed to the onsite
energy dierence of C sublattices. To gain more insights into this
phenomenon, it is worthwhile to investigate the charge density
dierence (CDD) induced by the BC3 substrate, expressed as
Dr = r(G/BC3)  r(G)  r(BC3),

(3)

where r(G/BC3), r(G), and r(BC3) are the total charge densities
of the G/BC3 hybrid structure, isolated graphene, and BC3
monolayer, respectively. All stacking patterns exhibit a very
similar CDD feature, thus we only present the P-II configuration.

Fig. 3

Fig. 4 Plane-integrated electron density dierence of P-II. The inset

shows the 3D isosurface of the charge density dierence, the yellow and
cyan areas represent electron accumulation and depletion, respectively.

As shown in the inset in Fig. 4, there is no charge accumulation in

the interspace between BC3 and graphene. However, in the graphene layer, there is evident charge redistribution by forming
electron-rich and hole-rich regions due to the dierent electrostatic
potentials induced by BC3 layer, which has also been found in the
G/BN HBLs.14 According to the Bader charge analysis, the redistribution makes little charge (B0.12e) transfer from graphene to the
BC3 substrate. However, the traces of charge transfer cannot drop
the Fermi level which still lies in the band gap as shown in Fig. 2(a).
It induces an intrinsic electric field, which points from the BC3 layer
towards graphene, and thus a sizable band gap opens at the Dirac
point. In addition, we also calculate the plane-integrated electron
density dierence to visualize the electron redistribution upon the
formation of the interface in Fig. 4. The electron rearrangement
localized at the interspace is visually revealed in the inset of Fig. 4.

The calculated total and partial DOS for P-II of G/BC3 HBLs.

This journal is the Owner Societies 2014

Phys. Chem. Chem. Phys., 2014, 16, 22861--22866 | 22863

View Article Online

Published on 04 September 2014. Downloaded by University of California - Riverside on 23/04/2015 05:43:09.

Paper

Fig. 5 The variation of band gap as a function of interlayer spacing for P-II.
(a) and (b) are the band structures of P-II at dierent spacing. The black
square and red ring represent the band gap of the K point and entirety.

The trend of change in the band gap of P-II is further

investigated by varying the interlayer distance between graphene
and BC3 monolayer, as shown in Fig. 5. When the distance
decreases, the charge begins to accumulate in the interspace of
graphene and BC3 monolayer. Compared with the equilibrium
state, more charge transfers from graphene to the BC3 substrate.
It means that there exists very strong hybridization between
graphene and BC3 monolayer. Therefore, the onsite energy
dierence of graphene increases and thus results in a larger
band gap opening. With the distance increasing, the interaction
between graphene and BC3 becomes weak. In this case, the
van der Waals interaction will gain the main role in the system.
The variation range is 14584% of distance-free gap from 2.4 to
3.4 . Similar results are also found for other configurations.
By examining the Dirac point position near the Fermi level,
however, unexpected semiconductormetal transition is observed
when the interlayer spacing is smaller than 2.5 or larger than
3.4 [see the inset of Fig. 5]. Beyond the range of [2.5, 3.4] ,
although the gap is still opened at the K point, an interlayerinduced self-doping occurs due to the movement of bands at the K
point, which significantly reduces the gap for the whole band
structure and renders graphene n or p-type metallic behaviors.
To further visualize the above theoretical model, we describe
the CDD with dierent interlayer spacing as shown in Fig. 6.
Owing to similar CDD features exhibited in all patterns, we just
plot the P-II configuration as an example. As per the above
analysis, with the decrease of the interlayer spacing, the charge
accumulation begins to occur between graphene and the BC3
substrate, also the G/BC3 HBL is still a semiconductor with a
band gap of 0.294 eV when the spacing is 2.6 . However, for
the case of d = 2.5 [Fig. 6(a)], it can be observed that a strong
orbital hybridization exists in the interspace of HBL, with a
large charge transfer from graphene to the substrate. Then it
forms hole-rich regions, which can improve the Fermi level of
graphene, leading to a metal state. When the distance is
between 2.6 and 3.5 , all the CDD features are similar to
the case of the equilibrium state [Fig. 6(b)]. For d = 3.6 , it can

22864 | Phys. Chem. Chem. Phys., 2014, 16, 22861--22866

PCCP

be obviously seen that no charge transfer and accumulation

occurs between graphene and BC3 layer. However, the larger
distance does not annihilate the eect of BC3 monolayer and
the defect-substrate induces a larger dipole moment in C atoms
which are located above the B atoms. This may be the main
reason for forming the phenomenon of electron-rich regions in
G/BC3 HBLs. Hence, these results indicate that the band
structure of graphene is sensitive to the interlayer spacing
between graphene and BC3 monolayer, providing an intuitive
theory model. However, tuning the interlayer spacing of bilayer
structures in experiments is still an awkward question with a
great challenge. So we will introduce one suggestion for achieving
the distance variation. Experimentally, the interlayer spacing of
bilayer structures is usually modulated by inserting small organic
molecules (such as NH4+) into them, which is considered in the
stripping of boron nitride and bilayer graphene. So, we expect
that this way can be used to tune the interlayer spacing of
G/BC3 HBLs.
The band gap of G/BC3 HBLs can also be further tuned by
the strain which can be easily achieved in experiments, as
shown in Fig. 7. The homogeneous in-plane strains are applied
to the HBLs by changing the lattices as e = (a  a0)/a0, where a
(a0) is the lattice constant under the strain (equilibrium)
conditions. Under the tensile strains, the graphene in G/BC3
HBLs preserves the planar plane clearly with the longer CC
bond length in the horizontal direction, also the variations of
the band gap for all HBLs exhibit the same trend. As is
mentioned above, the C-2p orbitals in graphene dominate the
state around the Fermi level. When starting to implement a
tensile strain with e = 1%, the band gap will decrease due to the
stretched CC bonds, which will weaken the p state between the
C atoms, leading to the energy shifting and the charge redistribution decrease. However, the continuous increase of tensile
strain and CC bond lengths will enhance the localization of
the electronic state on C atoms in graphene, resulting in the
increase of the band gap. Furthermore, we also investigate the
band gap change in the case of compressive strains. When
the compressive strains are small enough, graphene in all
G/BC3 HBLs can still maintain its planar structure. Nevertheless, it will obtain corrugation (B0.6 ) when the compressive
strains increase to 4%, and the BC3 substrate also gains
buckling (B0.3 ). For the variation of the band gap, complicated trends can be observed which are sensitive to the stacking
patterns. It does not increase monotonically as found in G/BN
or G/WS2 HBLs. In the range of [1%, 3%], the band gap
presents a descending trend due to the average of the electronic
states between neighboring C atoms, which will result in the
decrease of the onsite energy difference. On the other hand, the
following change in the band gap will be complex, which is
mainly attributed to the structural distortion of graphene and
the substrate induced by compressive strains. To reduce the
deformation of graphene and the substrate, experimentally, the
biaxial strain can be achieved by depositing G/BC3 HBLs on an
elastic substrate. In addition, in the process of tuning the band
gap by strain, the phenomenon of self-doping can still be
observed in Fig. 7. So this suggests that the strain plays an

This journal is the Owner Societies 2014

View Article Online

Published on 04 September 2014. Downloaded by University of California - Riverside on 23/04/2015 05:43:09.

PCCP

Fig. 6

Paper

Charge density dierence for the P-II stacking pattern at the interlayer spacing of d = 2.5, 3.3, 3.5 , respectively.

important role in tuning the band gap of graphene, which can

provide a viable route for the design of graphene-based functional nanostructures.
To realize the practical applications of G/BC3 HBLs in FETs,
higher carrier mobility and linear band dispersion are very
essential. So we investigate the electron eective mass (me*)
and the hole eective mass (mh*) at the Dirac point of G/BC3
HBLs. As we have mentioned before, the G/BC3 HBLs preserve
the linear band dispersion relationship of graphene. According
to the graphene dispersion relationship, the me* and mh* can
be expressed as2731
m 

Fig. 7 The variation of band gap as a function of strain for all stacking patterns,
where positive(negative) values of strain represent tensile(compressive) strains.
The red points represent the gap in the K point with self-doping. The inset
shows the schematic diagram.

p 
hk

vg vF

(4)

where k is the wave vector and vF is the Fermi velocity. Table 1

presents the calculated fitting value of vF for all G/BC3 HBLs.
Fortunately, it can be seen that all the stacking patterns attain
a considerable vF value (B0.76  106 m s1), which is

Table 1 The calculated binding energy per C atom (Eb), lattice constants (a0), interlayer distance (d0), energy band gap (Eg), eective masses of electrons
and holes, and the maximum Fermi velocity (vF) for all stacking patterns

Pattern

Eb (eV)

a0 ()

d0 ()

Eg (eV)

Eective mass

GK

KM

VF (m s1)

P-I

0.095

2.565

3.24

0.141

me*
mh*

6.75  1019
6.85  1019

6.14  1019
6.3  1019

0.77  106

P-II

0.099

2.565

3.29

0.162

me*
mh*

7.26  1019
7.38  1019

6.47  1019
6.85  1019

0.73  106

P-III

0.097

2.565

3.34

0.109

me*
mh*

6.47  1019
6.65  1019

6.22  1019
6.56  1019

0.76  106

This journal is the Owner Societies 2014

Phys. Chem. Chem. Phys., 2014, 16, 22861--22866 | 22865

View Article Online

Published on 04 September 2014. Downloaded by University of California - Riverside on 23/04/2015 05:43:09.

Paper

comparable to that of FSG (1  106 m s1). Moreover, according

to eqn (4), we also investigate the calculated me* and mh* along
GK and KM directions of all the stacking patterns, as shown
in Table 1. It can be observed that the me* and mh* at the
VBM and CBM of the G/BC3 HBLs are very small, which is
still more superior to other graphene-based nanoelectronics.
This indicates that the G/BC3 HBLs may have a higher carrier
mobility which is necessary in the application of FETs. The
large value of vF and the small value of m* are indispensable for
the design of applications and this can be analyzed qualitatively
by the interaction between graphene and the BC3 substrate. The
small binding energies indicate the weakening of electronic
coupling between graphene and BC3 and enable the graphene
to preserve linear band dispersion similar to that of FSG,
suggesting that the vF in this case may be comparable to that
of graphene. As a result, an ideal FET with the high carrier
mobility and a certain band gap on G/BC3 HBL based nanoelectronics can be expected to be achieved.

IV. Conclusion
A comprehensive first principles study is performed on the
structural and electronic properties of G/BC3 HBLs. Three
possible stacking patterns have been considered. The calculated band gaps of dierent stacking patterns are considerably
larger than kBT (26 meV) at room temperature; this may overcome
one of the main obstacles to apply graphene as an electronic
device, namely, the lack of intrinsic band gaps. We also find that
the interlayer spacing and in-plane strain can tune the band gap
of G/BC3 HBLs eectively. Interestingly, the characteristics of a
Dirac cone with a nearly linear band dispersion relationship of
graphene can be preserved, accompanied by a small electron
eective mass, and thus the carrier mobility is expected not to
degrade much. These findings may stimulate the experimental
developments in the graphene/BC3 HBLs with small energy gaps
and enable their use in novel integrated functional nanodevices.

Acknowledgements
This work was supported by National Natural Science Foundation of China (Grant no. 11274143, 60471042 and 11304121).

References
1 K. S. Novoselov, A. K. Geim, S. Morozov, D. Jiang, Y. Zhang,
S. Dubonos, I. Grigorieva and A. Firsov, Science, 2004, 306, 666.
2 Y. Zhang, Y.-W. Tan, H. L. Stormer and P. Kim, Nature, 2005,
438, 201.
3 C. Berger, Z. Song, X. Li, X. Wu, N. Brown, C. Naud,
D. Mayou, T. Li, J. Hass, A. N. Marchenkov, E. H. Conrad,
P. N. First and W. A. de Heer, Science, 2006, 312, 1191.
4 P. Avouris, Nano Lett., 2010, 10, 4285.

22866 | Phys. Chem. Chem. Phys., 2014, 16, 22861--22866

PCCP

5 Y.-W. Son, M. L. Cohen and S. G. Louie, Phys. Rev. Lett.,

2006, 97, 216803.
6 Y.-W. Son, M. L. Cohen and S. G. Louie, Nature, 2006,
444, 347.
7 V. Barone, O. Hod and G. E. Scuseria, Nano Lett., 2006,
6, 2748.
8 M. Ezawa, Phys. Rev. B: Condens. Matter Mater. Phys., 2006,
73, 045432.
9 M. Ishigami, J. H. Chen, W. G. Cullen, M. S. Fuhrer and
E. D. William, Nano Lett., 2007, 7, 16431648.
10 N. T. Cuong, M. Otani and S. Okada, Phys. Rev. Lett., 2011,
106, 106801106804.
11 S. Y. Zhou, G.-H. Gweon, A. V. Fedorov, P. N. First, W. A.
de Heer, D.-H. Lee, F. Guinea, A. H. Castro Neto and
A. Lanzara, Nat. Mater., 2007, 6, 770775.
12 E. Rollings, et al., J. Phys. Chem. Solids, 2006, 67, 21722177.
13 Y. Ma, Y. Dai, W. Wei, C. Niu, L. Yu and B. Huang, J. Phys.
Chem. C, 2011, 115, 2023720241.
14 S. S. Li and C. W. Zhang, J. Appl. Phys., 2013, 114, 183709.
15 Y. Fan, M. Zhao, Z. Wang, X. Zhang and H. Zhang, Appl.
Phys. Lett., 2011, 98, 083103.
16 H. Yanagisawa, T. Tanaka, Y. Ishida, E. Rokuta, S. Otani and
C. Oshima, Phys. Rev. B: Condens. Matter Mater. Phys., 2006,
73, 045412.
17 Y. Ding, Y. Wang and J. Ni, Appl. Phys. Lett., 2009,
94, 073111.
18 X. Chen and J. Ni, Phys. Rev. B: Condens. Matter Mater. Phys.,
2013, 88, 115430.
19 G. Kresse and J. Furthmuller, Phys. Rev. B: Condens. Matter
Mater. Phys., 1996, 54, 11169.
20 G. Kresse and J. Hafner, Phys. Rev. B: Condens. Matter Mater.
Phys., 1994, 49, 14251.
21 G. Kresse and J. Furthmuller, Phys. Rev. B: Condens. Matter
Mater. Phys., 1996, 54, 11169.
22 J. P. Perdew and Y. Wang, Phys. Rev. B: Condens. Matter
Mater. Phys., 1992, 45, 13244.
23 S. Grimme, J. Comput. Chem., 2006, 27, 1787.
ngya
`gue and J. G. A
n, J. Phys.
24 T. Bucko, J. Hafner, S. Lebe
Chem. A, 2010, 114, 11814.
25 G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter
Mater. Phys., 1994, 59, 1758.
26 H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Condens. Matter
Mater. Phys., 1976, 13, 5188.
27 V. Ariel, arXiv:1205.3995v1, 2012.
28 V. Ariel and A. Natan, arXiv:1206.6100, 2012.
29 K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, M. I.
Katsnelson, I. V. Grigorieva, S. V. Dubonos and A. A. Firsov,
Nature, 2005, 438, 197.
30 P. Avouris, Z. Chen and V. Perebeinos, Nat. Nanotechnol.,
2007, 2, 605615.
31 A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov
and A. K. Geim, Rev. Mod. Phys., 2009, 81, 109.

This journal is the Owner Societies 2014