strongly and weakly interacting configurations of...
TRANSCRIPT
STRONGLY AND WEAKLY INTERACTINGCONFIGURATIONS OF HEXAGONAL BORON
NITRIDE ON NICKEL
ABBAS EBNONNASIR*, SUNEEL KODAMBAKA*,‡,¶
and CRISTIAN V. CIOBANU†,§,¶
*Department of Materials Science and Engineering,University of California Los Angeles, Los Angeles, California 90095, USA
†Department of Mechanical Engineering and Materials Science Program,Colorado School of Mines, Golden, Colorado 80401, USA
‡[email protected]§[email protected]
Received 7 March 2015Revised 20 July 2015Accepted 23 July 2015
Published 4 September 2015
Using density functional theory calculations with van der Waals corrections, we have investigatedthe stability and electronic properties of monolayer hexagonal boron nitride (hBN) on the Ni(111)surface. We have found that hBN can bind either strongly (chemisorption) or weakly to the sub-strate with metallic or insulating properties, respectively. While the more stable con¯guration is thechemisorbed structure, many weakly bound (physisorbed) states can be realized via growth aroundan hBN nucleus trapped in an o®-registry position. This ¯nding provides an explanation forseemingly contradictory sets of reports on the con¯guration of hBN on Ni(111).
Keywords: Hexagonal boron nitride; nickel; graphene; density functional theory.
With the recent advances in controllable growth of
large-area, high-quality, two-dimensional (2D) lay-
ered materials (e.g. graphene1–4 and hexagonal boron
nitride (hBN)5,6) on metallic substrates, the ¯eld has
naturally progressed towards the design and/or
investigation of heterostructures based on out-of-
plane7–12 or in-plane13–15 stacking of these nanoma-
terials. Layered heterostructures of graphene on
boron nitride o®er the best avenue of preserving the
exotic properties of graphene and using them in
nanoscale devices, because of the near-perfect epi-
taxial lattice match, small van der Waals (vdW)
interactions between graphene and hBN, and most
importantly because of the insulating nature of the
underlying hBN layer(s). While only recently mono-
layer graphene been grown on7 or transferred8 onto
hBN, the e®orts to grow graphene/hBN hetero-
structures on metallic substrates date back more than
a decade.16–23
In particular, Oshima's group used chemical vapor
deposition (CVD) to grow hBN on a Ni(111) sub-
strate, followed by the CVD of graphene on top of
hBN/Ni(111) with benzene as precursor.21 Experi-
mental reports note that the CVD-deposited hBN
is often chemically bound to the Ni substrate, with
a 1� 1 structure and a metallic character.21–23
¶Corresponding authors.
Surface Review and Letters, Vol. 22, No. 6 (2015) 1550078 (6 pages)°c World Scienti¯c Publishing CompanyDOI: 10.1142/S0218625X1550078X
1550078-1
Independent experimental groups con¯rm the struc-
ture and metallic properties of hBN/Ni,24 both of
which are also qualitatively consistent with density
functional theory (DFT) calculations in the local
density approximation (LDA).25 However, there is
also signi¯cant evidence of weakly bound (physi-
sorbed) hBN on Ni(111),16–19 although the con¯gu-
ration of weakly bound hBN is not clear. The reason
why both physisorbed and chemisorbed hBN on
Ni(111) can be synthesized under very similar exper-
imental conditions is presently not well understood.
Naturally, in the physisorbed con¯gurations the vdW
interactions play a determinant role and their con-
sideration within the DFT framework is necessary.
In order to understand the seemingly contradic-
tory reports of physisorbed16–19 and chemisorbed21–23
hBN on Ni(111), we have used DFT calculations that
account for vdW interactions to determine the bind-
ing and electronic structure of hBN/Ni(111), and
have studied the binding energy as a function of the
interplanar distance from the hBN layer to the top
Ni(111) atomic plane. In the absence of any misori-
entation (strict epitaxy, no moir�e patterns), we ¯nd
that a monolayer of hBN on Ni(111) has two local
energy minima. The deeper minimum corresponds to
strong chemical bonding with the substrate, with the
N atoms located directly above ¯rst-layer Ni atoms;
the shallower minimum corresponds to a weaker
bound con¯guration in which the boron atoms are
directly above the top layer Ni atoms. This ¯nding
reconciles the experiments showing chemically bond-
ed hBN20–23 with those reporting physisorbed hBN
monolayers on Ni(111).16–19 Furthermore, in order to
assess the possibility that graphene may debond a
strongly bound hBN layer from the Ni substrate (as
proposed, e.g. in Refs. 21 and 26) we have also studied
the binding energy and electronic properties of hBN/
Ni in the presence of a graphene layer above hBN. We
have found that the interaction between graphene
and the rest of the system is practically independent
on the con¯guration of hBN on the substrate, and
that the electronic properties of hBN (in either the
weakly or the strongly bound con¯guration) remain
virtually the same in the presence of the graphene
layer.
Our spin-polarized, zero total charge DFT calcu-
lations27 were performed in the generalized-gradient
approximation (GGA) using the vdW functional of
Dion et al.28 as parameterized by Klimeš et al.29 This
functional performs very well in reproducing the
structural details and electronic properties of vdW
bonded materials involving hBN and graphene.30 We
used non-relativistic Troullier–Martins pseudopoten-
tials,31 double-� polarization (DZP) functions, and
a mesh-cuto® of 200Ry. The Brillouin zone was
sampled with 30� 30� 1 and 50� 50� 1 �-centered
k-point grids for structural relaxations and band
structure calculations, respectively; convergence with
respect to k-point density was ensured separately for
structure and electronic properties. All supercells are
1� 1 and have a slab geometry with the graphene
and/or hBN layers present only on one side of the six-
layer Ni(111) substrate, and with a vacuum thickness
of 15�A. The residual atomic forces were smaller than
0.03 eV/�A after the structural relaxations. With these
parameters, the calculated in-plane lattice constants
were 2.560�A for Ni, 2.512�A for hBN and 2.485�A for
graphene. We have used the lattice constant of hBN,
2.512�A, for all supercells. The reason for this choice is
that straining the substrate32 rather than the gra-
phene or hBN monolayers ensures that the e®ects of
registry with respect to the substrate are not obfus-
cated by strain.
We ¯rst describe our results concerning the
monolayer hBN on Ni(111). We investigated all
possible con¯gurations in which the in-plane forces on
all atoms are zero by symmetry, and several starting
con¯gurations for which they are not. In the former
case, we found that the only strongly bound con¯g-
uration is the N-top one, consistent with other stud-
ies34; for the latter case of arbitrary in-plane starting
forces, the DFT relaxations push the system either in
the N-top con¯guration, or into a physisorbed one.
Therefore, we describe here only two con¯gurations
for hBN on Ni(111): N-top and B-top, in which N and
B atoms, respectively, are directly above ¯rst-layer Ni
atoms, as shown in Fig. 1. If the starting con¯gura-
tion of this system is N-top, then the ¯nal state after
relaxation remains in this N-top registry, while be-
coming atomically corrugated (rumpled), with the
plane of B atoms lying 0.12�A below that of the N
atoms; the mean distance from the hBN layer to the
substrate is 2.17�A. On the other hand, for the B-top
con¯guration, the relaxed structure is virtually °at
and resides 3.01�A away from the substrate. These
GGAþ vdW results are in line with previous LDA
results25 as far as the interlayer distance and the
¯nal con¯guration are concerned. However, there are
A. Ebnonnasir, S. Kodambaka & C. V. Ciobanu
1550078-2
quantitative di®erences, one of which being the larger
value for the binding between hBN and the substrate
in our system due the inclusion of vdW interactions.
The total energy of each con¯guration relative to the
energy of the unbound system (i.e. in which the hBN
layer and the substrate are far apart) is shown in
Fig. 1, as a function of the mean interlayer distance;
this is the same as binding energy, but with the
negative sign retained so as to depict the familiar
well-pro¯le characteristic of bound states. For the
N-top case shown in Fig. 1, the value of the rumpling
is ¯xed at 0.12�A, i.e. at the value corresponding to
the relaxed structure. The N-top con¯guration with
the hBN layer kept atomically °at is also shown in
Fig. 1. Despite such constraint, the N-top °at con-
¯guration is also strongly bound to the substrate. The
binding energies and the structural parameters of the
hBN/Ni system are summarized in the ¯rst two rows
of Table 1. Based on the interlayer distance, the N-top
and B-top con¯gurations described in Fig. 1 corre-
spond to chemical and physical bonding, respectively.
The chemical bonding is strong and leads to a com-
mensurate structure of hBN/Ni. On the contrary, the
physisorbed hBN may not be unique, and a variety of
weakly-bound con¯gurations other than B-top, com-
mensurate, 1� 1 structure can be expected.
We note that the interlayer distance is not the sole
indication of how strong the binding with the sub-
strate is. Charge transfer at the interface is another
indicator of the binding strength. Indeed, as shown
in Fig. 2, the N-top con¯guration determines large
electronic density transfer at the interface, consistent
with strong, chemical bonding. On the same scale, the
electron transfer density for the B-top con¯guration is
signi¯cantly smaller than that in the N-top case:
coupled with the large interlayer spacing (of the order
of interlayer spacing in graphite), this indicates weak
physical binding for the B-top con¯guration.
As we have already noted, both the physi-
sorbed16–19 and chemisorbed20–22 con¯gurations were
observed experimentally, even though the latter is
signi¯cantly more favorable energetically (Table 1).
Based on the results in Fig. 1, we infer here why the
physisorbed state survives to withstand experimental
observation and characterization.16 From the mini-
mum of the B-top curve (Fig. 1) to its intersection
with the N-top (corrugated) curve, there is an energy
barrier to climb of �b ¼ 6meV/atom. Only if such
average barrier per atom is somehow overcome, could
the system break away from the physisorbed mini-
mum and fall into the deeper, chemisorbed one. We
assume that temperature T is the only provider of
energy variations and that an n-atom hBN cluster
Fig. 1. (Color online) Binding energy of an hBN mono-layer on Ni(111), for two di®erent registries (shown asinsets): Nitrogen on top of a ¯rst-layer Ni atom (solid cir-cles, con¯guration named N-top) and boron on top of a¯rst-layer Ni atom (triangles, B-top). We also consideredthe binding of a °at hBN layer with Ni(111) (dashed curve).
Table 1. Binding energies, interlayer distances and atomic corrugation (rum-pling) values for hBN/Ni(111) and gr/hBN/Ni(111).
Binding Interlayer Rumpling
energy (eV/atom) distance (�A) (�A)
hBN-Ni gr-hBN/Ni hBN-Ni gr-hBN hBN gr
hBN/Ni (N-top) 0.29 2.17 0.12hBN/Ni (B-top) 0.16 3.01 0.03gr/hBN/Ni (N-top) 0.29 0.17 2.33 3.05 0.05 0.00gr/hBN/Ni (B-top) 0.16 0.16 3.07 3.11 0.01 0.00
Strongly and Weakly Interacting Con¯gurations of hBN on Nickel
1550078-3
remains °at. Under this assumption, the presence of a
barrier (�b ¼ 6meV/atom) implies that if the n-atom
nucleus happens to be in the physisorbed local mini-
mum, then it will remain trapped in that local mini-
mum as long as the thermal energy kBT (where kB is
the Boltzmann constant) is not su±cient to take it
past the barrier �b per atom, i.e. as long as
kBT � n�b. The larger the size n of the nucleus is, the
harder it is to be taken away from the physisorbed
state by temperature alone. As a rough estimate, at
processing temperatures21 of T ¼ 1070K, a nucleus of
hBN containing more than kBT=�b � 15 atoms
would be large enough to survive and grow into a
monolayer domain in the physisorbed state. Given
that often hBN is grown via CVD of borazine, this is
only about three six-atom hBN rings in the minimum
nucleus. This being said, there remains the question
of why the nucleation may happen in the physisorbed
state, when the chemisorbed state is markedly more
stable. This is because there are numerous physi-
sorbed states (most of them incommensurate) in
which the system might \take o®", but there is only
one chemisorbed one (i.e. the N-top inset of Fig. 1).
As a cautionary note, we note that these arguments
are not likely to hold at the scale of large hBN sheets,
since at such scales ripples can form33 a®ecting their
planarity; further studies may be necessary in the
future to elucidate the details of registry transitions
for monolayer sheets and °akes.
In order to assess the possibility that an additional
graphene layer on the hBN/Ni system may alter the
electronic properties of hBN and its binding to the
substrate, we turn our attention to the graphene/
hBN/substrate system in which graphene is placed
above the hBN layer in a Bernal con¯guration. We
have also studied other di®erent relative registries of
the graphene monolayer with respect to hBN/Ni,
with the same conclusions. The results are summa-
rized in the last two rows of Table 1. As seen in the
table, the graphene layer does not a®ect much the
binding of the hBN to the substrate, neither in the
B-top nor in the N-top con¯guration. The binding
energy of graphene to hBN/Ni is signi¯cant, 0.16–
0.17 eV/atom, and is due to the dispersion forces.
However, graphene does not detach the hBN layer
from the substrate, but rather adjusts itself at a dis-
tance of � 3.1�A from the hBN layer. In the N-top
con¯guration, the B atoms are pulled out a bit,
resulting in a decreased corrugation of the hBN sheet
(Table 1). Contrary to an early proposal suggesting
that this corrugation decrease leads to the debonding
of hBN from the substrate,35 we have found that the
hBN moves only slightly away from the substrate and
does not debond. Furthermore, as we have shown in
Fig. 1 (the curve labeled \N-top °at"), even if rum-
pling were set to zero, the binding of hBN to the
substrate remains very strong, comparable to that in
the N-top corrugated con¯guration.
When analyzing the electronic properties, in par-
ticular band structure, we ¯nd that in the N-top
con¯guration the hBN layer is metallic (Fig. 3), as
there are �-d hybridized bands corresponding to hBN
and Ni that cross the Fermi level. The hybridized
bands shown in Fig. 3 are indicative of chemical
bonding between hBN and the substrate, and are
nearly identical to those of the hBN/Ni system
without graphene; this shows that graphene does not
change the electronic structure at the hBN/Ni inter-
face. We have also analyzed the band structure
(Fig. 4) of the gr/hBN/Ni system with hBN in the B-
top con¯guration and nearly equally spaced between
graphene and the nickel substrate (refer to last row of
Table 1). As shown in Fig. 4, the hBN layer is now
insulating with a large band gap (� 4 eV). The band
structure shown in Fig. 4 for the B-top con¯guration
is very similar to that obtained for the B-top hBN/Ni
N
Ni
B
B
N
Ni
B
(a)
(b)
Fig. 2. (Color online) Charge transfer density at the hBN/Ni(111) interface for (a) the N-top and (b) the B-top con-¯gurations. The color scale is the same for both pictures.
A. Ebnonnasir, S. Kodambaka & C. V. Ciobanu
1550078-4
system in the absence of the graphene layer. Based on
the electronic structure analysis, and on the rather
small structural changes induced by graphene
(Table 1), we conclude that the presence of a gra-
phene layer on top of hBN/Ni does not signi¯cantly
change the properties of the hBN/Ni interface irre-
spective of the registry and binding (chemical or
physical) that hBN has with the substrate. Since the
mere presence of the graphene layer does not alter the
electronic properties of hBN/Ni in N-top or B-top
registries (much less change the binding21,23 between
hBN and the Ni substrate from chemical to physical),
it follows that a more likely explanation for the
debonding reports of Oshima et al.21,23 should stem
from unintended processes happening during the
CVD growth of graphene on hBN/Ni.
In conclusion, we have shown that hBN has two
adsorption con¯gurations on Ni(111), one chemical
and one physical. Even though the chemically bonded
hBN is more energetically favorable, the weakly
bound can also be achieved when a small cluster is
trapped in the B-top con¯guration, acting as a nu-
cleus around which the physisorbed layer can grow.
This explains the seemingly contradictory reports of
strongly and weakly bound hBN on Ni from experi-
ments with similar parameters of the CVD growth.
We have also found that the presence of a graphene
layer on top of hBN/Ni does not signi¯cantly alter
the binding and the electronic properties of the hBN/
Ni interface. The presence of two adsorption states
with the same crystalline orientation with respect to
the substrate but with markedly di®erent electronic
properties (one metallic, one insulating), while inter-
esting in itself, could ¯nd potential applications in
nanoscale electromechanical switching devices that
toggle between the two states when stimulated
externally.
Acknowledgments
The authors gratefully acknowledge funding from
National Science Foundation (A. E. and C. V. C)
through Grant No. CMMI-0846858 and from the
O±ce of Naval Research (A. E. and S. K.) through
ONR Award No. N00014-12-1-0518 (Dr. Chagaan
Baatar). Supercomputer time for this project was
provided by the Golden Energy Computing Organi-
zation at Colorado School of Mines.
References
1. X. S. Li, W. W. Cai, J. H. Ann, S. Kim, J. Nah, D. X.Yang, R. Piner, A. Velamakanni, I. Jung, E. Tutuc,
Fig. 4. (Color online) Electronic structure of the gra-phene/hBN/Ni(111) heterostructure, with the hBN layer inthe B-top registry. The � and d character of the bands areindicated by the colors (inset legend). The structure of hBNis now physisorbed and insulating, with a band gap greaterthan 4 eV at the K-point.
Fig. 3. (Color online) Electronic structure of the gra-phene/hBN/Ni(111) heterostructure, with the N-top reg-istry for hBN/Ni and graphene Bernal-stacked on hBN.The � and d character of the bands are indicated by thecolors (inset legend). The structure of hBN is metallic, sincepz states of B and N are present across the Fermi level — asshown both in the band structure and in projected densityof states (PDOS). In this con¯guration, hBN is chemicallybonded to Ni (N-top con¯guration), at a distance of 2.17�A.
Strongly and Weakly Interacting Con¯gurations of hBN on Nickel
1550078-5
S. K. Banerjee, L. Colombo and R. S. Ruo®, Science324 (2009) 1312.
2. S. Bae, H. Kim, Y. Lee, X. F. Xu, J. S. Park, Y. Zeng,J. Balakrishnan, T. Lei, H. R. Kim, Y. I. Song, Y. J.Kim, K. S. Kim, B. Ozyilmaz, J. H. Ahn, B. H. Kongand S. Iijima, Nature Nanotechnol. 5 (2010) 574.
3. Z. Z. Sun, Z. Yan, J. Yao, E. Beitler, Y. Zhu and J. M.Tour, Nature 468 (2010) 549.
4. A. Reina, S. Thiele, X. T. Jia, S. Bhaviripudi, M. S.Dresselhaus, J. A. Schaefer and J. Kong, Nano Res. 2(2009) 509.
5. K. H. Lee, H. J. Shin, J. Lee, I. Y. Lee, G. H. Kim, J. Y.Choi and S. W. Kim, Nano Lett. 12 (2012) 714.
6. L. Song, L. Ci, H. Lu, P. B. Sorokin, C. Jin, J. Ni, A. G.Kvashin, D. G. Kvashin, J. Lou, B. I. Yakobson and P.M. Ajayan, Nano Lett. 10 (2010) 3209.
7. W. Yang, G. Chen, Z. Shi, C.-C. Liu, L. Zhang, G. Xie,M. Cheng, D. Wang, R. Yang, D. Shi, K. Watanabe, T.Taniguchi, Y. Yao, Y. Zhang and G. Zhang, NatureMater. 12 (2013) 792.
8. J. M. Xue, J. Sanchez-Yamagishi, D. Bulmash, P.Jacquod, A. Deshpande, K. Watanabe, T. Taniguchi,P. Jarillo-Herrero and B. J. Leroy, Nature Mater. 10(2011) 282–285.
9. M. Wang, S. K. Jang, W. J. Jang, M. Kim, S. Y. Park,S. W. Kim, S. J. Kahng, J. Y. Choi, R. S. Ruo®, Y. J.Song and S. Lee. Adv. Mater. 25 (2013) 2746.
10. Z. Liu, L. Song, S. Z. Zhao, J. Q. Huang, L. L. Ma, J. N.Zhang, J. Lou and P. M. Ajayan, Nano Lett. 11 (2011)2032.
11. R. Cort�es, D. P. Acharya, C. V. Ciobanu, E. Sutterand P. Sutter, J. Phys. Chem. C 117 (2013) 20675.
12. A. Ramasubramaniam, D. Naveh and E. Towe, NanoLett. 11 (2011) 1070.
13. L. Ci, L. Song, C. Jin, D. Jariwala, D. Wu, Y. Li, A.Srivastava, Z. F. Wang, K. Storr, L. Balicas, F. Liuand P. M. Ajayan, Nature Mater. 9 (2010) 430.
14. P. Sutter, R. Cort�es, J. Lahiri and E. Sutter, NanoLett. 12 (2012) 4869.
15. S. Jun, X. Li, F. Meng and C. V. Ciobanu, Phys. Rev.B. 83 (2011) 153407.
16. A. Nagashima, N. Tejima, Y. Gamou, T. Kawai and C.Oshima, Phys. Rev. Lett. 75 (1995) 3918.
17. A. Nagashima, N. Tejima, Y. Gamou, T. Kawai and C.Oshima, Phys. Rev. B 51 (1995) 4606.
18. A. Nagashima, N. Tejima, Y. Gamou, T. Kawai, M.Terai, M. Wakabayashi and C. Oshima, Int. J. Mod.Phys. B 10 (1996) 3517.
19. A. Nagashima, N. Tejima, Y. Gamou, T. Kawai and C.Oshima, Surf. Sci. 357–358 (1996) 307.
20. E. Rokuta, Y. Hasegawa, K. Suzuki, Y. Gamou and C.Oshima, Phys. Rev. Lett. 79 (1997) 4609.
21. C. Oshima, A. Itoh, E. Rokuta, T. Tanaka, K.Yamashita and T. Sakurai, Solid State Commun. 116(2000) 37.
22. C. Oshima, N. Tanaka, A. Itoh, E. Rokuta, K.Yamashita and T. Sakurai, Surf. Rev. Lett. 7 (2000)521.
23. T. Tanaka, A. Ito, A. Tajima, E. Rokuta and C.Oshima, Surf. Rev. Lett. 10 (2003) 721.
24. W. Auwarter, T. J. Kreutz, T. Greber and J. Oster-walder, Surf. Sci. 429 (1999) 229.
25. M. N. Huda and L. Kleinman, Phys. Rev. B. 74 (2006)075418.
26. T.Kawasaki,T. Ichimura,H.Kishimoto,A.A.Akbar,T.Ogawa and C. Oshima, Surf. Rev. Lett. 9 (2002) 1459.
27. J. M. Soler, E. Artacho, J. D. Gale, A. Garcia, J.Junquera, P. Ordejón and D. S. Sǎnchez-Portal, J.Phys., Condens. Matter 14 (2002) 2745.
28. M. Dion, H. Rydberg, E. Schr€oder, D. C. Langreth andB. I. Lundqvist, Phys. Rev. Lett. 92 (2004) 246401.
29. J. Klimeš, D. R. Bowler and A. Michaelides, Phys. Rev.B 83 (2011) 195131.
30. G. Graziano, J. Klimeš, F. Fernandez-Alonso and A.Michaelides, J. Phys., Condens. Matter 24 (2012)424216.
31. N. Troullier and J. L. Martins, Phys. Rev. B 43 (1991)1993; ibid Phys. Rev. B 43 (1991) 8861.
32. G. Giovannetti, P. A. Khomyakov, G. Brocks, V. M.Karpan, J. van den Brink and P. J. Kelly, Phys. Rev.Lett. 101 (2008) 026803.
33. S. K. Singh, M. Neek-Amal, S. Costamagna and F. M.Peeters, Phys. Rev. B 87 (2013) 184106.
34. G. B. Grad, P. Blaha, K. Schwarz, W. Auwarter andT. Greber, Phys. Rev. B 68 (2003) 085404.
35. C. Oshima and A. Nagashima, J. Phys., Condens.Matter 9 (1997) 1.
A. Ebnonnasir, S. Kodambaka & C. V. Ciobanu
1550078-6