Sie sind auf Seite 1von 8

This content has been downloaded from IOPscience. Please scroll down to see the full text.

Download details:
IP Address: 200.159.241.37
This content was downloaded on 19/08/2014 at 20:32
Please note that terms and conditions apply.
Group-IV nanosheets with vacancies: a tight-binding extended Hckel study
View the table of contents for this issue, or go to the journal homepage for more
2014 J. Phys.: Condens. Matter 26 365501
(http://iopscience.iop.org/0953-8984/26/36/365501)
Home Search Collections Journals About Contact us My IOPscience
Journal of Physics: Condensed Matter
J. Phys.: Condens. Matter 26 (2014) 365501 (7pp) doi:10.1088/0953-8984/26/36/365501
Group-IV nanosheets with vacancies:
a tight-binding extended H uckel study
Adriano de Souza Martins and Marcos Verssimo-Alves
Departamento de Fsica, Instituto de Ci encias Exatas, Universidade Federal Fluminense, Niter oi, RJ,
Brazil
E-mail: marcos verissimo@id.uff.br
Received 27 March 2014, revised 23 June 2014
Accepted for publication 8 July 2014
Published 18 August 2014
Abstract
In this work, we present a theoretical study of the electronic properties of group-IV element
nanosheets, namely graphene, silicene, germanene and the corresponding hydrogenated
structures for the two latter, silicane and germanane. We compare the results of two different
calculation methods, Density Functional Theory (DFT) and Extended H uckel Theory (EHT),
for both pristine sheets and sheets of silicene and germanene with a single-atom vacancy. We
show that EHT offers a remarkably reliable description of the electronic structure of these
materials for all cases, thus offering an affordable way for studying large systems for which
DFT calculations would be expensive and lengthy.
Keywords: electronic structure, extended H uckel method, density functional theory,
tight-binding, nanostructures
(Some gures may appear in colour only in the online journal)
1. Introduction
Since 2004, graphene has attracted great interest because of
its unique electronic properties, such as a linear dispersion in
its band structure near the Dirac point. The synthesis of fully
hydrogenated graphene, graphane, opened the possibility of
its application for hydrogen storage. Similar materials made
of Si and Ge are of extreme relevance, since integration into
current Si-based technology would be smoother: the recent
experimental characterization of silicene [1] as well as Density
Functional Theory (DFT) calculations [24] reveal buckled
sheets instead of at ones, reecting the reduced tendency of
sp
2
bonding for Group-IV elements.
Theoretical studies of the electronic properties of silicene
and germamene [4, 5] predict these materials to be gapless
semiconductors with linear energy dispersion relations near the
Kpoints, much like graphene. Fully hydrogenated silicene and
germanene, named silicane and germanane, respectively, have
been theoretically studied [4, 6] and both result to be wide-gap
materials from DFT calculations.
The understanding of changes to the electronic structure
in these materials due to the presence of defects, such as vacan-
cies, substitutional impurities or adsorbed atoms (adatoms) on
the surfaces is of utmost importance, not only with technologi-
cal applications in mind, but also for their use as a laboratory
for fundamental physics. So far, DFT [7] has been the method
of choice for addressing such changes, due to the reliability
of results obtained. Nevertheless, DFT is highly CPU- and
memory-demanding, limiting the theoretical study of systems
to unit cells containing up to a few hundreds of atoms. For
systems with defects like vacancies, small supercells intro-
duce spurious interactions between the defect and its periodic
images.
Semi-empirical calculations, such as tight-binding (TB)
methods, can be a good alternative to accurate but compu-
tationally expensive rst-principles calculations such as DFT.
By replacing the self-consistent process of the determination of
the interactions between atoms in the system by parametrized
Hamiltonians, substantially larger cells can be used with rela-
tively modest computational requirements. However, a perti-
nent question is: are dramatic changes to the electronic struc-
ture of materials due to the presence of defects properly taken
into account by semi-empirical methods?
Considering the vacancy defect, in [8] an interesting way
of quantifying the inuence of vacancies in the electronic
structure is offered: while maintaining the atoms in their
0953-8984/14/365501+07$33.00 1 2014 IOP Publishing Ltd Printed in the UK
J. Phys.: Condens. Matter 26 (2014) 365501 A de Souza Martins and M Verssimo-Alves
ideal positions, changes in the hopping integrals between two
nearest neighbours of a vacancy, which arise from the lifting
of the orthogonality condition between the orbitals of the
removed atom and the orbitals of the neighbouring atoms, are
introduced.
In this work, we present theoretical calculations based on
the Extended H uckel Theory (EHT) of the band structure for
pristine graphene, silicene, germanene and the corresponding
hydrogenated structures for the two latter systems. In addition,
in order to assess the reliability of the H uckel method for
dealing with a non-ideal structure, the electronic structure
of silicene and germanene with single-atom vacancies is
calculated with EHT, and the results are compared with plane-
wave basis set DFT calculations.
The article is organized as follows. In section 2, a sum-
mary of the main ideas underlying the most common theoret-
ical approaches to the electronic structure of nanostructures
is presented, followed by a brief summary of the Extended
H uckel Theory and the method for obtained the parameters of
the EHT calculations in section 2.1. In section 3.1, the re-
sults of the parametrization of the H uckel Hamiltonian from
DFT calculations are presented. In section 3.2 the electronic
structure of the group-IV nanosheets, both pristine and with
vacancies, fromDFTand H uckel calculations using the param-
eters determined in section 3.1, are presented and discussed.
Finally, in section 4 we present the conclusions drawn from
our results.
2. Methodology
Since the seminal work of Slater and Koster [9], the tight-
binding (TB) formalism has been a reliable tool for describing
the electronic properties of crystalline solids. In the orthog-
onal formulation of the TB method, the crystalline states are
described as linear combinations of atomic orbitals (LCAO),
where the basis functionsthe (valence) atomic orbitals for
each chemical speciesare assumed to constitute an orthonor-
mal set. In other words, the basis orbitals of a given atom do
not overlap with those fromthe other atoms of the system. This
assumption allows atomic orbitals to be expressed in terms of
known functions, which are used as formal tools to construct
all matrix elements of the Hamiltonian.
Empirical, orthogonal tight-binding (OTB) methods are
quick and practical and its parameters, namely orbital energies
and hoppings, are adjusted in order to reproduce the band
structure of the target material in a specic geometry. Because
of that, OTB parameters in the orthogonal formulation
are usually not very transferable to different chemical
environments and, in addition, their values should be adjusted
if the system suffers structural deformations. Tight-binding
basis sets are commonly assumed to be both orthogonal and
short ranged, while atomic wave functions are not, meaning
that OTBbasis sets donot resemble the eigenstates of anatomic
Hamiltonian. Improvements in the OTB band structures can
be obtained through the inclusion of hoppings beyond rst-
nearest-neighbours, and through the addition of more orbitals
tothe basis set. For systems withstrain, where atoms donot rest
in the ideal crystalline positions, a common way to circumvent
this problem is to correct hopping parameters using Harrisons
power-law scaling [10], but this approach is successful only
for a small range of strain values.
For the systems treated in this article, the orthogonal TB
schemes do not result in a very accurate description of the un-
derlying band structures, as discussed in [3] for silicene and sil-
icon nanotubes: there are noticeable discrepancies between the
TB and ab-initio bands along the MK line, even when a set
of TBparameters including hoppings up to next-nearest neigh-
bours is employed. This shows that OTB parameters have re-
duced transferability when the atomic environments changes,
a very undesirable feature since it requires re-parametrization
for each different material under consideration.
First-principles approaches, on the other hand, yield ac-
curate descriptions of band structures, and structural defor-
mations are naturally captured by total energy rst-principles
calculations by allowing the structure to relax to the equilib-
rium positions while determining electronic charge densities
self-consistently, fully taking into account the interplay of ge-
ometry and electronic structure. The process of self-consistent
determination of the electronic charge density is the most ex-
pensive aspect inmost total energyrst-principles calculations,
and a great deal of work is devoted to reduce the computational
burden required to carry out DFT calculations, such as the use
of O(N) methods for the case of strictly localized basis sets, in
which case the Hamiltonian matrices become sparse (as done,
for example, in the SIESTAcode [11]). As in the OTBmethod,
in the EHT method the process of self-consistently calculat-
ing the electronic charge density is avoided entirely, making it
faster and more affordable than DFT calculations.
2.1. The Extended H uckel Theory
The Extended H uckel Theory (EHT) is a semi-empirical
technique lying between the OTB and rst-principle limits:
the method works with explicit analytical expressions for
the basis orbitals. For a given geometry, one uses the
explicit EHT basis functions to calculate a non-orthogonal
overlap matrix S, which along with separately tted on-site
Hamiltonian matrix elements, yields the corresponding off-
diagonal hopping elements of the Hamiltonian. In the EHT
framework, valence orbitals are spanned in terms of a double
zeta Slater-Type Orbital (STO) basis [12, 13], and structural
changes are simply accounted for by recalculating the overlap
and hopping elements, but leaving the basis sets and onsite
elements unchanged
1
.
In our work, we use the same method described in [12, 13],
which we briey summarize below. Using a double-zeta spd
STO basis set, we explicitly calculate the overlap matrix S:
S

=
_

d
3
r (1)
and the elements of the Hamiltonian matrix, given by
H

= E

=
1
2
K
EHT
_
H

+ H

_
S

, (2)
1
This approach yields good results only in the case of moderate structural
deformation, which would justify not recalculating on-site energies. See [13].
2
J. Phys.: Condens. Matter 26 (2014) 365501 A de Souza Martins and M Verssimo-Alves
where K
EHT
is an additional tting parameter whose value is
commonly set to 1.75 for molecules and 2.3 for solids [12],
with reasonable exibility in the choice of these parameters.
The parametrizations for around 40 elements of the periodic
table can be found in [12], including crystalline silicon and
carbon.
For each of the chemical elements present, the
wavefunctions are expanded in the double-zeta spd basis set
according to [12]:

nl
=
2

i=1
c
i
r
n1
e

i
r
Y
lm
(, ) (3)
Therefore, each atom type requires the determination of 12
parameters: three onsite energies (E
s
, E
p
and E
d
), the zetas
and the rst expansion coefcient c
1
. The value of the c
2
coefcient is constrained so as to guarantee the normalization
of the orbital, while the remaining 11 parameters are generated
through a simulated annealing (SA) approach using the
algorithm proposed by Vanderbilt [14]. In this approach, the
H uckel parameters are varied in successive Monte Carlo cycles
with decreasing temperatures in order to reduce the value of
an objective function y, namely the root mean square (RMS)
deviation of the H uckel bands E
H
i
(k) respective to a target
band structure E
T
i
(k):
y =
1
nb nk

_
nb

i=1
nk

j=1
_
E
H
i
(k
j
) E
T
i
(k
j
)
_
2
, (4)
where nb and nk denote, respectively, the number of bands and
k-points.
In our work, the target band structures to which EHT
parameters are adjusted are those of pristine unit cells
of the group-IV nanosheets, fully optimized through DFT
calculations as input parameters for the SA tting procedure.
The electronic bands of all systems were generated along the
MKM lines, and the gaps for silicane and germanane
were further corrected by shifting the DFTconduction bands to
match the GW band gap [6]. In the SA procedure, all valence
bands and the two rst conduction bands were included in
equation (4) and the resulting t is considered to be acceptable
when y 0.15 eV.
Finally, the calculation of the tight-binding bands is
carried out within a non-orthogonal SlaterKoster scheme
[9]. The analytical formulas of Barnett [15] for , and
components of the overlap between the basis orbitals are
employed to calculate the overlap and Hamiltonian matrices S
and H. Within the non-orthogonal scheme, the band structure
of a system is obtained as the solution of a generalized
eigenvalue problem:
H(k)
i
(k) = E
i
(k)S(k)
i
(k), (5)
where
i
(k) denotes the eigenvector of the ith band, and
k is the Bloch wave vector within the rst Brillouin Zone.
Hoppings were restricted to sites with inter-atomic distances
smaller than 9 (cutoff radius).
Figure 1. Schematic side view of an innite silicene nanosheet. The
buckling parameter, , is indicated for clarity.
3. Results and discussion
3.1. DFT structures and H uckel parameters for pristine
group-IV nanosheets
It is well-known that the group-IV nanosheets treated here are
more stable in a buckled conguration which, as mentioned
in section 1, reects a reduced tendency of elements from
that group to form sp
2
bonds. It is also well-established that
silicene, germanene and their hydrogenated counterparts are
only stable in a low-buckling conguration, which is therefore
the only one for which we perform calculations. Figure 1
shows schematically a side view of the supercell for silicene,
indicating the buckling parameter, , which is dened as the
vertical distance between two nearest-neighbor atoms.
Table 1 shows calculation parameters, such as plane-wave
energy cutoffs and Brillouin Zone k-point samplings, and re-
sults of full structural optimization, which were used as input
for the SA tting procedure. In all DFT calculations, carried
out with the Quantum Espresso code [16], ultrasoft pseudopo-
tentials [17] were employed for the description of valence elec-
tronsof all chemical elements, using the Generalized Gradients
Approximation (GGA) of Perdew et al [18] as the exchange
and correlation energy functional. A dense MonkhorstPack
mesh [19] and a large vacuum layer were used to ensure high-
lyaccurate results. Plane-wave cutoffs for both wavefunc-
tion and charge densities yielded an energy convergence of
at least 10
4
Ry/atom in the case of non-hydrogenated struc-
tures, and 10
3
Ry per Si/H (Ge/H) group. For atomic relax-
ation, the convergence criteria was that all forces were lower
than 10
5
Ry/Bohr. The electronic DOS were calculated with
both EHT and DFT using a Gaussian smearing of 50 meV. The
results obtained are in excellent agreement with those of [20].
Tables 2 and 3 present the H uckel parameters for both
pristine (graphene, silicene and germanene) and hydrogenated
(silicane and germanane) group-IV nanosheets, all obtained
following the SA approach cited in section 2.1. The best
ts were obtained considering K
EHT
= 2.8 for graphene and
K
EHT
= 2.3 for the silicon- and germanium-based systems.
In all calculations, the Fermi level was xed at 13.0 eV.
Although all AO are spanned in a double- basis, there are
cases in which the AO is well described with just one Slater
orbital. For those cases, we used
2
= 25, and the value of
the c
2
coefcient, which ensures AOnormalization, is given as
c
2
=
_
1 c
2
1
. This means that we neglect the overlap among
the Slater basis orbital and their neighbours.
3
J. Phys.: Condens. Matter 26 (2014) 365501 A de Souza Martins and M Verssimo-Alves
Table 1. Parameters used in plane-wave calculations and results for all structures studied in this work. For all structures, the k-point
sampling is done with an 18 18 1 MonkhorstPack grid.
Graphene Silicene Germanene Silicane Germanane
Plane-wave cutoff energy (Ry) 28.0 38.0 38.0 38.0 38.0
Charge density cutoff energy (Ry) 200.0 290.0 380.0 270.0 255.0
MonkhorstPack grid 18181 18181 18181 18181 18181
Lattice constant, a () 2.463 3.867 4.043 3.886 4.073
Vacuum layer () 19.0 19.551 19.319 16.28 16.83
Atom-atom bond length () 1.422 2.277 2.431 2.357 2.464
Atom-H bond length () 1.500 1.553
Buckling distance () 0.0 0.450 0.680 0.720 0.734
Note: In the table, atom denotes the group-IV element of the nanosheet.
Table 2. Optimized parameters of the atomic orbitals (AO) basis set calculated by SA. The value of the c
2
coefcient is not included
whenever
2
= 25 (see text). The K value in equation (2) was set to 2.8 for graphene, and 2.3 for the silicene and germanene.
Atom AO
1
c
1

2
c
2
Graphene C 2s 1.889 0.561
2p 1.455 0.368 3.091 0.930
3d 0.948 0.498
Silicene Si 3s 1.803 0.636
3p 1.621 0.711 7.792 0.704
3d 0.966 0.468 9.890 0.884
Germanene Ge 4s 1.892 0.472
4p 1.925 0.797 9.324 0.604
4d 1.233 0.547 9.038 0.837
Silicane Si 3s 1.821 0.673 9.919 0.739
3p 1.482 0.518 2.689 0.855
3d 0.796 0.409 4.176 0.913
H 1s 1.070 0.591 10.048 0.807
Germanane Ge 4s 2.316 0.771
4p 1.715 0.533 2.804 0.846
4d 1.901 0.612 9.272 0.791
H 1s 1.070 0.591 10.048 0.807
Table 3. Optimized on-site energies, calculated with structural parameters from DFT calculations as listed in table 1.
Compound Element E
s
E
p
E
d
Graphene C 21.493 13.517 1.510
Silicene Si 18.668 12.491 5.142
Germanene Ge 19.256 11.115 5.872
Silicane Si 17.348 10.839 5.343
H 13.856
Germanane Ge 18.380 10.141 4.586
H 13.640
The values of the on-site energies depend on the chemical
environment, which in turn depends on the systems geome-
try. The optimum values for the 2D materials considered in
this work are summarized in table 3. In contrast with bulk
compounds [12], the transferability of the AO parameters to
different environments (such as silicene and silicane) was only
partial, as it can be noticed fromtable 2. The exception was the
Hatom, whose AOparameters were kept equal in both silicane
and germanane.
The resulting band structures of ve parametrized pristine
compounds are shown in gures 2 and 3. The parameters
determined from the SA procedure are seen to yield band
structures in very good agreement with DFT results.
3.2. Vacancies in silicene and germanene
In order to assess the reliability of the calculated H uckel param-
eters in different chemical environments, an electronic density
of states calculation (DOS) for silicene and germanene with
a single-atom vacancy was performed. The initial congura-
tions correspond to a 6 6 supercell with one missing atom,
amounting to 71 atoms relaxed by means of a DFT calculation
until forces were lower than 3 10
4
Ry/Bohr. The DOS was
then calculated in both DFT and EHT formalisms. In EHT
calculations, the DOS was calculated using the relaxed DFT
geometry, with the EHT parameters for the pristine cells listed
in tables 2 and 3.
4
J. Phys.: Condens. Matter 26 (2014) 365501 A de Souza Martins and M Verssimo-Alves
Figure 2. Band structure of pristine non-hydrogenated group-IV nanosheets. (a) Band structure of graphene. (b) Band structure of silicene.
(c) Band structure of germanene. Dots correspond to the target DFT GGA calculations. Solid lines correspond to the calculated H uckel
band structure. The Fermi level is indicated by a dashed line.
Figure 3. Band structure of pristine hydrogenated group-IV nanosheets. (a) Band structure of silicane. (b) Band structure of germanane.
Dots correspond to the target DFT GGA calculations, with the conduction bands rigidly shifted in order to match the GW band gaps of [6].
Solid lines correspond to the calculated H uckel band structure. The Fermi level is indicated by a dashed line.
Figure 4 shows the relaxed structure of silicene when a
single-atom vacancy is present. For this kind of vacancy,
in which two atoms form a weak bond, and one forms no
bonds, a strong deviation from equilibrium positions of the
pristine sheet can be noticed. In particular, two atoms form
a weak bond, and one is dangling. In a recent article, Gao et
al [21] studied vacancies in silicene and found that two kinds
of vacancies are present: the one we study in this work, which
they denote SV-2, and another, which they denote SV-1, in
which the three nearest-neighbours of the missing Si atomform
a chemical bond, restoring the original threefold coordination
at the expense of an increase in elastic energy. They nd that, in
spite of this increase inelastic energy, the SV-1vacancyis much
more stable than the SV-2 vacancy. Nevertheless, the study of
the SV-2 vacancy poses a much more challenging situation to
EHT calculations since the SV-2 vacancy is a charged defect.
Therefore, the SV-2 is a good test to the robustness of EHT
calculations withparameters determinedfor pristine structures.
Figures 57 show the DFT and EHT DOS for the
relaxedsupercells withthe single-atomvacancies for graphene,
silicene and germanene, respectively. From the DFT DOS
shown in gures 57, all structures with single-atomvacancies
display sharp peaks close to the Fermi level. These levels are
associated with localized electronic states due to the p-electron
Figure 4. Silicene with a single-atom vacancy after geometry
relaxation. Two of the atoms around the vacancy move to form a
weak bond, and the remaining atom is dangling, thus constituting a
charged defect. A similar geometry is found for germanene.
localized on the dangling group-IV atom. The EHT and DFT
DOS for all materials are remarkably similar, indicating that
even if the parameters determined for pristine structures are
used, EHTis still capable of capturing the physics of electronic
5
J. Phys.: Condens. Matter 26 (2014) 365501 A de Souza Martins and M Verssimo-Alves
0
10
20
30
40
50
-12 -10 -8 -6 -4 -2 0 2
D
O
S

[
S
t
a
t
e
s
/
e
l
e
c
t
r
o
n
]
Energy [eV]
DFT
Extended Huckel
Figure 5. DOS calculation for graphene with a single-atom vacancy.
Full (dashed) lines are results from DFT (EHT) calculations. The
Fermi level is set to 0 eV.
0
20
40
60
80
100
120
140
160
-12 -10 -8 -6 -4 -2 0 2
D
O
S

[
S
t
a
t
e
s
/
e
l
e
c
t
r
o
n
]
Energy [eV]
DFT
Extended Huckel
Figure 6. DOS calculation for silicene with a single-atom vacancy.
Full (dashed) lines are results from DFT (EHT) calculations. The
Fermi level is set to 0 eV.
0
20
40
60
80
100
120
140
160
-12 -10 -8 -6 -4 -2 0 2
D
O
S

[
S
t
a
t
e
s
/
e
l
e
c
t
r
o
n
]
Energy [eV]
DFT
Extended Huckel
Figure 7. DOS calculation for germanene with a single-atom
vacancy. Full (dashed) lines are results from DFT (EHT)
calculations. The Fermi level is set to 0 eV.
localization at a quantitative level, rather than at a simply
qualitative level. This means that a small calculation effort
is necessary to obtain the parameters needed to be used in
subsequent calculations in large systems where atoms can have
different coordination numbers.
In our EHT calculations for vacancies, we use the
relaxed structures from DFT calculations, which in the case
of non-hydrogenated structures, involved 71 atoms. If the
hydrogenated structures (silicane and germanane) were to be
calculated, this number would double, and the computational
effort would more than double, since plane-wave calculations
entail order-(N
3
) operations for the diagonalization of the
large Hamiltonian matrices involved. Given that the electronic
structure of these compounds is dependent on the relaxed
geometries even for DFT itself, it would be natural and fair
to ask if EHT, which in principle is not capable of performing
structural relaxations, presents any advantage in its use.
Tight-binding and EHT have recently been used in
electronic conduction studies of graphene nanoribbons with
topological defects [22, 23]. This choice was probably
motivated by the lower computational cost of electronic
transport calculations using this method, compared to the
Non-Equilibrium Greens Function Method, which is also the
method of choice for such calculations, but can demand rather
robust computational capabilities, depending on the choice of
Hamiltonian and the use of self-consistency in the calculation.
In both works, an initial geometry relaxation is performed with
DFT, TB/EHT parameters are then tted and subsequently the
electronic conduction is calculated in the TB/EHT framework.
This is because geometry relaxation in defective materials can
dramatically alter its electronic properties. From our results,
the use of EHT would be very advantageous for the study
effects of defect concentration and disorder on the electronic
structure of a given material, since it could reach numbers of
atoms larger by more than two orders of magnitude compared
to DFT calculations, with a relatively modest computational
effort. However, if geometrical relaxation effects are not taken
into account, then the method is of somewhat limited utility.
Recent work has been performed using the REAXFF
classical potential [24] for the structure of vacancies in silicene
when different numbers of atoms are removed [25], with
good agreement on structural results from DFT and classical
potential calculations. This means that geometries optimized
with appropriate classical potentials could be used as input
for EHT calculations, providing an affordable means to reach
supercells with large numbers of atoms with only reasonably
modest computational resources. Since EHTcan yield reliable
bandstructures once the correct equilibriumgeometryis found,
a combination of geometry relaxation using classical potentials
followed by EHT calculations could allow for faster screening
of strategies for tailoring of materials so as to achieve certain
desired properties. In fact, a proposal of similar spirit has been
made as early as 1978 [26], in the context of energy evaluation
of different conformers of aromatic molecules aiming to nd
ground state geometries.
4. Summary
In conclusion, we have studied the application of Extended
Huckel Theory (EHT) to two-dimensional group-IVmaterials,
with and without hydrogenation. Our investigation focused on
the agreement of band structures of EHT calculations with
those of DFT calculations, which are currently the method
of choice for electronic structure calculations. In particular,
we wanted to determine if EHT is sufciently robust as to
yield correct electronic structures using parameters obtained
for pristine unit cells of the materials under consideration.
Our results have a remarkably good agreement with results
6
J. Phys.: Condens. Matter 26 (2014) 365501 A de Souza Martins and M Verssimo-Alves
of complex and resource-demanding DFT calculations, which
is very encouraging.
To the best of our knowledge, this is the rst time an EHT
calculation has been carried out for charged defects in solid-
state systems. The excellent agreement between the electronic
structure obtained from EHT and DFT calculations should
open very interesting possibilities for areas in which accurate
and complex electronic structure methods are computationally
prohibitive, such as the evaluation of effects of disorder of
defects on electronic properties of materials.
Acknowledgments
The authors are grateful for the nancial support of the
Brazilian funding agency Fundac ao de Amparo ` a Pesquisa
do Estado do Rio de Janeiro (FAPERJ) through grant E-
26/112.554/2012.
References
[1] Vogt P, Padova P D, Quaresima C, Avila J, Frantzeskakis E,
Asensio M C, Resta A, Ealet B and Lay G L 2012 Silicene:
compelling experimental evidence for graphenelike
two-dimensional silicon Phys. Rev. Lett. 108 155501
[2] Takeda K and Shiraishi K 1994 Theoretical possibility of stage
corrugation in Si and Ge analogs of graphite Phys. Rev. B
50 14916
[3] Guzm an-Verri G G and Voon L C L Y 2007 Electronic
structure of silicon-based nanostructures Phys. Rev. B
76 075131
[4] Voon L C L Y, Sandberg E, Aga R S and Farajian A A 2010
Hydrogen compounds of group-IV nanosheets Appl. Phys.
Lett. 97 163114
[5] Houssa M, Pourtois G, Afanasev V V and Stesmans A 2010
Can silicon behave like graphene? A rst-principles study
Appl. Phys. Lett. 97 112106
[6] Houssa M, Pourtois G, Afanasev V V and Stesmans A 2011
Electronic properties of hydrogenated silicene and
germanene Appl. Phys. Lett. 98 223107
[7] Payne M C, Teter M P, Allan D C, Arias T A and
Joannopoulos J D 1992 Iterative minimization techniques
for ab initio total-energy calculations: molecular dynamics
and conjugate gradients Rev. Mod. Phys. 64 1045
[8] van der Rest J and Pecheur P 1983 Tight-binding calculation
of orbital relaxations and hopping integral modications
around vacancies and antisite defects in GaAs J. Phys.
France 44 1297
[9] Slater J C and Koster G F 1954 Simplied LCAO method for
the periodic potential problem Phys. Rev. 94 1498
[10] Capaz R B, de Ara ujo G C, Koiller B and von der Weid J P
1993 Pressure and composition effects on the gap properties
of Al
x
Ga
1x
As J. Appl. Phys. 74 5531
[11] Soler J M, Artacho E, Gale J D, Garca A, Junquera J,
Ordej on P and S anchez-Portal D 2002 The SIESTA method
for ab initio order-N materials simulation J. Phys.:
Condens. Matter 14 2745
[12] Cerd a J and Soria F 2000 Accurate and transferable
extended H uckel-type tight-binding parameters Phys. Rev. B
61 7965
[13] Ribeiro I A, Ribeiro F J and Martins A S 2014 An extended
H uckel study of the electronic properties of III-V
compounds and their alloys Solid State Commun. 186 50
[14] Vanderbilt D and Louie S G 1984 A Monte Carlo simulated
annealing approach to optimization over continuous
variables J. Comput. Phys. 56 259
[15] Barnett M P 2003 Molecular integrals and information
processing Int. J. Quantum Chem. 95 791
[16] Giannozzi P et al 2009 QUANTUM ESPRESSO: a modular
and open-source software project for quantum simulations
of materials J. Phys.: Condens. Matter 21 395502
[17] Vanderbilt D 1990 Soft self-consistent pseudopotentials
in a generalized eigenvalue formalism Phys. Rev. B
41 7892
[18] Perdew J P, Burke K and Ernzerhof M 1996 Generalized
gradient approximation made simple Phys. Rev. Lett.
77 3865
[19] Monkhorst H J and Pack J D 1976 Special points for
Brillouin-zone integrations Phys. Rev. B 13 5188
[20] Wang X Q, Li H D and Wang J T 2012 Induced
ferromagnetism in one-side semihydrogenated silicene and
germanene Phys. Chem. Chem. Phys. 14 3031
[21] Gao J, Zhang J, Liu H, Zhang Q and Zhao J 2013 Structures,
mobilities, electronic and magnetic properties of point
defects in silicene Nanoscale 5 9785
[22] Ihnatsenka S and Kirczenow G 2013 Effect of edge
reconstruction and electronelectron interactions on
quantum transport in graphene nanoribbons Phys. Rev. B
88 125430
[23] Dubois S M M, Lopez-Bezanilla A, Cresti A, Triozon F,
Biel B, Charlier J C and Roche S 2010 Quantum transport
in graphene nanoribbons: effects of edge reconstruction and
chemical reactivity ACS Nano 4 1971
[24] van Duin A C T, Dasgupta S, Lorant F and III W A G 2001
ReaxFF: a reactive force eld for hydrocarbons J. Phys.
Chem. A 105 9396
[25] Berdiyorov G R and Peeters F M 2014 Inuence of vacancy
defects on the thermal stability of silicene: a reactive
molecular dynamics study RSC Adv. 4 1133
[26] Dougherty D A and Mislow K 1979 A combination of
empirical force eld and extended H uckel molecular orbital
calculations as a computational approach to conformational
analysis J. Am. Chem. Soc. 101 1401
7

Das könnte Ihnen auch gefallen