Physics Letters A 375 (2011) 939945

Contents lists available at ScienceDirect

Physics Letters A
www.elsevier.com/locate/pla
Electronic structure and optical properties of substitutional and interstitial
phosphor-doped ZnO
Li Guan

, Baoting Liu, Qiang Li, Yang Zhou, Jianxin Guo, Guoqi Jia, Qingxun Zhao,
Yinglong Wang, Guangsheng Fu
College of Physical Science and Technology, Hebei University, Baoding 071002, Peoples Republic of China
a r t i c l e i n f o a b s t r a c t
Article history:
Received 7 May 2010
Received in revised form 24 December 2010
Accepted 26 December 2010
Available online 28 December 2010
Communicated by R. Wu
Keywords:
P-doped ZnO
Optical properties
Electronic properties
First-principles
Using rst-principles method, electronic structure and optical properties of phosphorus-doped ZnO for
the possible substitutional (P
Zn
, P
O
) and interstitial (P
tet
, P
oct
) doping are investigated. P
O
gives p-type
conductivity, but others show n-type. P
O
and P
tet
has a signicant difference in optical properties due to
the contribution of P 3p states at Fermi level.
2010 Elsevier B.V. All rights reserved.
1. Introduction
As a wide band gap semiconductor, ZnO has recently attracted
great attention because of potential applications in optical and op-
toelectronic devices, such as lasers [1], light emitting diodes [24],
transparent electrodes in solar cells [5] and photodetectors [6]. Ex-
perimentally, n-type ZnO thin lm with good conductive properties
can be easily obtained by doping impurities in pure ZnO [711].
However, it is dicult to fabricate p-type ZnO due to a large num-
ber of native defects such as Zn interstitials (Zn
i
) and/or O vacan-
cies (V
O
) [12]. At present, attractive acceptor-state candidates for
ZnO are group-V elements substituted at the O site. Several groups
have attempted to achieve p-type ZnO with group-V elements
using different experimental methods [1317]. Some researchers
found that N-doped ZnO could be unstable and dicult to be re-
peated [1820]. More recently, P monodoped [21,22], (PIn) [23]
and (PN) [24] codoped ZnO have realized p-type conductivity.
P-doped ZnO having the trends for both n- and p-type have been
observed in experiments [2527]. The lack of stable p-type ma-
terials is still the key issue which hinders the development of
ZnO-based optoelectronic devices [28,29]. To reveal the origins
of the characters, rst-principles calculations were performed to
study the electronic structures of native and P-related defects and
in pure and P-doped ZnO [2833]. Based on rst-principles cal-
*
Corresponding author.
E-mail address: lguan@hbu.edu.cn (L. Guan).
culations, the theory for large-sized-mismatched group-V dopants
[28] and a P
Zn
2V
Zn
complex contributing to p-type doping e-
ciency in P-doped ZnO [29] have been presented. More recently,
Tian et al. found that P
Zn
4N
O
complexes may have lower forma-
tion energy than that of P
Zn
under Zn-rich condition, which helps
improving the p-type conductivity in ZnO [33]. We have performed
rst-principles calculations to reveal the contribution of P
Zn
and
P
O
to the conductivity type of P-doped ZnO [34]. In fact, the pos-
sible positions (interstitial and substitutional) of P in ZnO should
be taken into account [35]. It is a pity that there is not more in-
formation about interstitial P-related defects in the previous rst-
principles calculations. Moreover, some groups have experimen-
tally investigated the optical properties of P-doped ZnO [25,36,37],
yet little microscopic information on the formation mechanism has
emerged [15].
In the present Letter, we report on a density functional study of
the electronic and optical properties of P-doped ZnO for the substi-
tutional (P
Zn
, P
O
) and interstitial (P
tet
, P
oct
) doping congurations.
We rstly examined the formation energies of the different doping
sites, and it has been found that P
O
is more dicult to achieve and
P
Zn
is energetically favorable. Secondly, according to band struc-
tures and densities of state of pure and P-doped ZnO, we found
that a substitutional P at an O site forms a deep acceptor level
and a substitutional P at a Zn site behaves as a donor. P atoms
at both tetrahedral and octahedral interstitial sites contribute to
n-type conductivity. Thirdly, the optical properties are determined
by the band structures, and the optical constants, including the
optical absorption, reectivity, refractive index, extinction coe-
0375-9601/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.physleta.2010.12.064
940 L. Guan et al. / Physics Letters A 375 (2011) 939945
Fig. 1. The supercell atomic structure of P-doped ZnO used in the calculations. Gray balls indicate Zn and red balls represent O. Here, P
O
and the interstitial site as the center
of tetrahedral (P
tet
), and P
Zn
and the octahedral interstitial site (P
oct
) are indicated in (a) and (b), respectively. (For interpretation of the references to color in this gure
legend, the reader is referred to the web version of this Letter.)
cient and energy loss function, have been discussed for pure and
P-doped ZnO. Compared with the other doping congurations, P
at O site and tetrahedral interstitial site will result in a remark-
able difference in optical properties, such as the optical absorption
intensity in visible-UV range.
2. Computational details
The calculations are performed with CASTEP (Cambridge serial
total energy package) code [38], which is based on the density
functional theory using a plane-wave pseudopotential method. This
method can be used to calculate the periodic structural mate-
rials. In this Letter, we used the PerdewBurkeErnzerhof (PBE)
[39] to describe the exchange-correlation function in the scheme
of generalized gradient approximation (GGA) and selected the ul-
trasoft pseudopotential (USP). O 2s
2
2p
4
, Zn 3d
10
4s
2
and P 3s
2
3p
3
as the valence-electron congurations were used to describe the
electronion interaction. We employed a supercell containing 36
host atoms in the wurtzite structure. The four dopant sites of P
atoms including substitutional and interstitial sites are indicated
as Fig. 1. In the optimization process, a plane wave basis set with
an energy cutoff of 340 eV was used to expand the electronic
wave functions. Supercell structures were fully relaxed and the
free movement of the atoms in any direction was allowed and no
symmetry constraint was imposed. In the procedure of calculation,
the atomic coordinates were optimized to achieve the energy min-
imization. A 4 4 2 Brillouin-zone k-point sampling was used in
the calculations. The convergence in energy, maximum force, max-
imum stress, and maximum displacement tolerances were set as
1.0 10
5
eV/atom, 0.03 eV/, 0.05 GPa, and 0.001 , respec-
tively.
The formation energies of the P dopants at substitutional and
interstitial dopant sites in ZnO are expressed respectively, by
E
f
(P
s
) = E
def
(P
s
) E
per
(ZnO) (
P

s
) (1)
E
f
(P
i
) = E
def
(P
i
) E
per
(ZnO)
P
(2)
where E
f
(P
s
) and E
f
(P
i
) indicate the formation energies of substi-
tutional and interstitial P defects, respectively. E
def
(P
s
) and E
def
(P
i
)
are the total energies of a ZnO supercell containing substitutional
and interstitial P defects, respectively. E
per
(ZnO) is the total energy
of a perfect supercell of ZnO.
P
is the phosphor chemical po-
tential, and
s
(s = Zn and O) denotes the corresponding atomic
chemical potential. For ZnO, the chemical potential
O
=
1
2
E(O
2
)
and
Zn
= E(Zn) is given by the total energy per each Zn atom
in bulk zinc. H
bulk
(ZnO) is the formation entropy of bulk ZnO cal-
culated as 4.02 eV/ZnO in our calculations, which is compared to
the experimental value of 3.63 eV [40] and the previous theoretical
result of 3.94 eV [35].
3. Results and discussion
3.1. Electronic properties of bulk ZnO
Band structure and partial density of state (PDOS) of bulk ZnO
have been calculated in our previous work [34]. Bulk ZnO is a di-
rect band gap semiconductor and its band gap is 0.72 eV, which is
larger than the theoretical result of X.D. Zhang et al. [41] (0.5 eV),
but much smaller than the experimental value of 3.37 eV [42]. It is
well known that the underestimated band gap is due to exchange-
correlation energy. Fig. 2 gives the imaginary part
2
() of dielec-
tric function of pure ZnO compared with the experimental values
of Refs. [43] and [44].
2
() of pure ZnO has two main peaks at
4.19 and 9.10 eV respectively. In this part, we have used the scis-
sor operation, which is a kind of method that has been proved [45,
46]. The scissors operator is a rigid shift and can describe the dif-
ference between the theoretical and experimental band gap values.
When the experimental value is known, you can perform a band
structure calculation by setting the scissors operator to the differ-
ence between the two values [47]. By GGA method, the calculated
peak of 4.19 eV is due to optical transitions between O 2p states
in the valence band maximum and Zn 4s states in the conduction
band minimum. The peak of 9.10 eV is caused by optical transi-
tions between Zn 3d states and O 2p states. The peaks are very
close to the other rst-principles estimations [41,48,49]. The above
L. Guan et al. / Physics Letters A 375 (2011) 939945 941
Fig. 2. The imaginary part
2
() of dielectric function of pure ZnO. (a) Our theo-
retical values by GGA method, (b) and (c) are the experimental values of Refs. [43]
and [44], respectively.
results indicate that our computational method is effective. Based
on the optimized geometry structure of pure ZnO, electronic struc-
tures and optical properties of P-doped ZnO will be calculated in
the followings.
3.2. Atomic and electronic properties of P-doped ZnO
The interatomic bonding properties of P-doped ZnO at the four
dopant sites are investigated. Comparing the ionic radii of the
P
3
(about 0.1800.212 nm), P
3+
(about 0.0440.058 nm), P
5+
(about 0.0310.034 nm), O
2
(about 0.1380.140 nm), Zn
2+
(about
0.074 nm) [5053], ZnO bond distance (0.200 nm) [29], and the
octahedron interstice radius (0.062 nm) of bulk ZnO [54], we pre-
dicted that P substituting O sites could be dicult due to the
increase of the lattice spacings. The calculated ZnO bond distance
of bulk ZnO is 0.201 nm, which is in agreement with the value
given in Ref. [29]. The calculated charge distributions of P impuri-
ties in ZnO indicate that P with negative charges (0.52e) at the
O sites, and positive charges (+1.25e, +0.26e and +0.58e) at the
Zn sites, and tetrahedral and octahedral interstitial sites, respec-
tively. A high overlap indicates a high degree of covalency in the
bond, and a value close to zero indicates that there is no signicant
interaction between the electronic populations of the two atoms
[55]. For P
Zn
, we can see that three PO bonds are formed when
P at Zn lattice site, and the Mulliken charge overlap population
(MCOP) of the neighboring OO bond is 0.11, which shows an
antibonding interaction between the two O atoms, that is to say,
the two ZnO bonds tend to be broken and one weak OO anti-
bonding state is formed. The interatomic bonding possibly implies
that the doping structure may tend to be a P
Zn
2V
Zn
complex un-
der a suitable condition predicted by Lees work [29]. In optimized
geometry structure of P
O
in ZnO, the PZn bond lengths are calcu-
lated to be 2.25 , which are larger than the calculated ZnO bond
length of 2.01 for bulk ZnO. The transferred charge of P
tet
to the
neighbour atoms is less than that of P
oct
. P
tet
O bonds (1.51 ) are
shorter than P
oct
O bonds (1.63 ), and then the MCOP of P
tet
O
bond (0.56) is larger than that of P
oct
O bond (0.37), indicating
a stronger interaction between P 3p and O 2s electrons for P
tet

ZnO. Therefore, we predicted that the charge distribution of P

tet
could contribute to its particular PDOS and then optical properties.
Based on Eqs. (1) and (2), it is estimated that the formation
energy of P
O
is the largest, and relative to P
O
, the formation en-
ergies of P
Zn
, P
tet
and P
oct
are 5.1 eV, 2.25 eV and 1.99 eV,
respectively. The results indicate that P at Zn lattice site is ener-
Fig. 3. Total density of state (a) pure ZnO; (b) P
Zn
ZnO; (c) P
O
ZnO; (d) P
tet
ZnO;
(e) P
oct
ZnO. The Fermi energy is indicated by the dashed vertical line. Here, the
results of (a)(c) are detailedly illustrated in Ref. [34].
getically favorable, and both the interstitial P
tet
and P
oct
are not
stable and easy to migrate from the original central positions. The
previously experimental works reported that P doping at Zn sites
signicantly increases the electron carrier density in as-deposited
lms, making the material heavily n-type, while the P atom seems
to occupy the O site in ZnO samples at higher annealing temper-
atures, making the material p-type [21,56,57]. However, the effect
of annealing temperature on the hole concentration in p-type ZnO
is not well understood [57]. To make the physical mechanism clear,
we calculated the total DOSs for P-doped ZnO indicated as Fig. 3.
It can be clearly seen that except P
O
ZnO as p-type conductivity,
the DOSs of other doped ZnO materials show n-type. A substitu-
tional P at an O lattice site forms a deep acceptor level of 1.02 eV
above the valence band maximum, which is in agreement with
the earlier calculation of 0.96 eV [33]. The interstitial P atoms in
ZnO behave as donors in our calculations, while the interstitial As
atoms in ZnO were regarded as the amphoteric point defects in
the previous work [28]. The calculated formation energies show
that P atoms preferentially substitute Zn sites, making the mate-
rial n-type. In Vaithianathans work of Ref. [21], they found that
after rapid thermal annealing at 600

C, P
O
is for p-type conductiv-
ity, but the as-deposited ZnO:P lms revealed n-type conductivity
with the use of Zn
3
P
2
. Using X-ray absorption near-edge structure
spectroscopy, they did not found that any spectral feature that can
be associated with P
Zn
2V
Zn
clusters for p-type conductivity. In
Ref. [57], the authors found that hole concentration was increased
with an increase in activation temperature in N
2
, Ar, and O
2
am-
bient gases. Their results indicated the P
O
rather than P
Zn
2V
Zn
to act as acceptors for the P
2
O
5
source. Associated with our cal-
culated results, the experimental measurements show that higher
activation energy is necessary to form P
O
, and p-type conductivity
of ZnO will be achieved with an increase in annealing temperature.
The conductivity type of P-doped ZnO material varies with the
dopant sites of P impurities, so the PDOSs of P atoms at four
dopant sites are illustrated in Fig. 4. The contributions of P atoms
to the density of states near the Fermi level are remarkably differ-
942 L. Guan et al. / Physics Letters A 375 (2011) 939945
Fig. 4. Partial density of state of P impurities in (a) P
Zn
ZnO; (b) P
O
ZnO; (c) P
tet

ZnO; (d) P
oct
ZnO. The Fermi energy is indicated by the dashed vertical line. The
results of (a)(b) are detailedly illustrated in Ref. [34].
ent. For P
Zn
, P 3p states enter into the conduction band, while no
obvious states appear near the Fermi level. For P
O
, new impurity
states mainly attribute to P 3p states in the bottom of forbidden
band and the Fermi level above the impurity energy level. The
analysis of charge transfer indicated that P impurity has negative
charges (0.52e) at the O sites. Therefore, it is concluded that
P
O
ZnO shows weak p-type conductivity. For P
oct
, there is a lower
density of impurity states at Fermi level compared to that of P
tet
.
For P
tet
, P 3p state is very close to the conduction band minimum
and shows a higher density at Fermi level, which strongly inu-
ences the optical transition between the valence and conduction
band.
3.3. Optical properties of P-doped ZnO
As we know, the imaginary part
2
() of the dielectric constant
is given by

2
() =
2e
2

k,v,c

c
k

ur

v
k
_

_
E
c
k
E
v
k
E
_
(3)
where u is the vector dening the polarization of the incident elec-
tric eld, is the light frequency, e is the electronic charge,
is the unit cell volume, and
c
k
and
v
k
are the conduction and
valence band wave functions, respectively.
2
() can be regarded
as detailing the real transitions between the occupied and unoccu-
pied electronic states. The real part of the dielectric function
1
()
can be evaluated from the imaginary part
2
() by the famous
KramersKronig relationship. The absorption coecient (), re-
ectivity R(), refraction index n(), and extinction coecient
k() can be obtained from
1
() and
2
() as the following ex-
pressions
() =

2
_
_

1
()
2
+
2
()
2

1
()
_
1/2
(4)
R() =

()
1 +()

2
(5)
n() =
1

2
_
_

1
()
2
+
2
()
2
+
1
()
_
1/2
(6)
k() =
()
2
(7)
Fig. 5. The imaginary part of the dielectric function of pure ZnO and P-doped ZnO.
The energy loss function is calculated directly from the total di-
electric function as
Im
_
1
()
_
=

2
()

1
()
2
+
2
()
2
(8)
To ensure numerical convergence, we calculated the optical
properties of P-doped ZnO with 72-atom supercell, which can
satisfy the convergence condition [29,33]. Fig. 5 gives the imag-
inary parts of the dielectric function spectra of P-doped ZnO by
GGA + U approach (U = 7.5 eV). The remaining gap underesti-
mation within the GGA + U approach is corrected by a scissor
operator of =1.51 eV. Compared with pure ZnO, all of the
2
()
spectra of doped ZnO exhibit peaks in the lower energy region
(<3 eV), where those of P
O
ZnO and P
tet
ZnO show much higher
peaks. For P
O
ZnO, O atoms substituted by P atoms leads to the in-
crease of the lattice spacings. Due to the valence bands of P
O
ZnO
moving up and the band gap decreasing, the peaks slightly shift to
the low energy range relative to undoped ZnO. A pronounced peak
originating from the impurity absorption locates at about 1.6 eV
in this spectrum and can be revealed explicitly from band struc-
ture shown in Fig. 6(a). P impurity band appears in the bottom
of the forbidden band and is close to the valence band maximum
(VBM). The most possible electron transitions indicated by the red
and green arrows in Fig. 6(a) are from P impurity bands at K point
to the conduction bands at the same point. The
2
() spectrum
of P
Zn
ZnO only has three main peaks appear at 1.98, 5.10 and
7.03 eV respectively. The threshold energy of absorption at about
0.9 eV is attributed to the electronic transitions between VBM and
conduction band minimum (CBM), and the rst peak is ascribed to
the interband transitions at K point, which are shown as the blue
and red arrow in Fig. 6(b), respectively. For P
oct
ZnO, P impurity
acts as a deep donor and the impurity band appears in the up-
per forbidden band. From the
2
() spectrum, it can be seen that
a broad peak at about 2.3 eV in the lower energy region, which
is associated with the transitions from the valence bands to the
impurity bands at K point indicated as the red arrow in Fig. 6(c).
The impurity band of P
tet
ZnO is very close to CBM and a highest
peak appears at about 1.6 eV due to the transition from the va-
lence bands to the impurity bands indicated by the red and green
arrows in Fig. 6(d). Therefore, we can conclude that the difference
between the imaginary parts of the dielectric function of P-doped
ZnO originates from the different contributions of P 3p states at
Fermi level.
Fig. 7 reveals the absorption coecient and reectivity spec-
tra of pure ZnO and P-doped ZnO. The absorption edge of bulk
L. Guan et al. / Physics Letters A 375 (2011) 939945 943
Fig. 6. The band structures of (a) P
O
, (b) P
Zn
, (c) P
oct
, and (d) P
tet
of P-doped ZnO. The arrows indicate electron transitions and the horizontal dot lines indicate the Fermi
level. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of this Letter.)
Fig. 7. The absorption coecient and reectivity spectra of pure and P-doped ZnO.
ZnO is 3.1 eV (400 nm), which ts very well with the experimen-
tal value 386 nm [58]. Compared to pure ZnO, P
Zn
ZnO shows
two main absorption peaks 7.83 eV and 14.32 eV in the high
energy range, and the absorption spectrum of P
oct
ZnO shows a
similar tendency to that of P
Zn
ZnO. For P
O
ZnO and P
tet
ZnO,
the absorption peaks slightly shift to the low energy range rela-
tive to that of pure ZnO and the pronounced absorption at lower
energies due to the electron transitions from P impurity band il-
944 L. Guan et al. / Physics Letters A 375 (2011) 939945
Fig. 8. The refractive index n() and extinction coecient k() of pure and P-doped ZnO.
lustrated as Figs. 6(a) and (d). From Fig. 7, it can be seen that
there are four obvious reectivity peaks at 3.27, 5.08, 6.78 and
9.13 eV of pure ZnO, and the experimental values are 3.10, 5.0,
8.65 and 12.7 eV, respectively [48]. In the higher energy region,
the discrepancy between the experimental measurements and our
calculated values increases, and this has also been found in the
previously theoretical work [59]. For P
O
ZnO and P
tet
ZnO, the
intensities of rst reectivity peaks are much higher than those
of other P-doped ZnO. The refractive index n() and extinction
coecient k() of pure and P-doped ZnO are shown in Fig. 8.
In the range of the photon energy lower than 3 eV, both n()
and k() show remarkable differences from other discussed cases,
attributing to the transitions from the impurity states in the for-
bidden band to conduction bands or valence bands. In Fig. 9, the
energy loss function describes an energy loss that when the elec-
trons get through the uniform dielectric medium and the loss
function peak exhibits the frequency of plasma resonance. This
function can be compared the theoretical results with spectroscopy
(e.g., electron energy loss spectroscopy) measurements which pro-
vide information about the electronic system of the sample in-
teracting with an incident electron beam. A signicant feature
in the low-loss spectrum (050 eV) is the bulk plasmon peak,
which has an intensity several orders of magnitude higher than
the core-loss edges due to collective oscillation of the loosely
bound electrons [60]. The highest peak in energy loss spectrum
corresponds to the dramatical descent of reectivity shown in
Fig. 7.
For P-doped ZnO, the optical properties in the high energy
range (>10 eV) are controlled by electrons transitions between O
2p states and Zn 3d states or O 2s states and Zn 3d states. There-
fore, the spectra tend to be identical in this wavelength range for
undoped and doped ZnO. However, the optical properties in the
low energy range are mainly affected by the electron transitions
from impurity bands. According to the calculated results displayed
as Figs. 69, in the visibleUV wavelength range, P atom at O site
or the tetrahedral interstitial site plays a signicant role on the
optical properties. Base on the PDOS of P
tet
, a higher density of
electronic states composed of P 3p exists at Fermi level leads to a
stronger optical transition. Thus, the peak intensities of P
O
ZnO
and P
tet
ZnO in visibleUV range are remarkably different from
those of the others. The pronounced properties of the two dop-
ing structures can give a new perspective of physical properties of
P-doped ZnO.
Fig. 9. The energy loss function of pure and P-doped ZnO.
4. Conclusions
In this Letter, we have investigated atomic and electronic struc-
ture and optical properties of pure ZnO and P-doped ZnO using
the ultrasoft pseudopotential based on density functional theory.
P atoms at substitutional and interstitial sites are considered in
our calculations and the results show that the formation energy of
P
O
is higher than those of others P-related defects, while P
Zn
is en-
ergetically favorable and thus Zn site acts as a preferential dopant
site in crystal lattice. Only P
O
as a deep acceptor contributes to
p-type conductivity in all of the discussed doping congurations.
Either P
O
or P
tet
results in a high density of P 3p electron states at
Fermi level, which signicantly affects its optical properties in the
visibleUV range.
Acknowledgements
This work was supported by National Natural Science Founda-
tion of China (Grant No. 10847116), Ph.D. Programs Foundation of
Ministry of Education of China (Grant No. 20091301110002), China
Postdoctoral Science Foundation (Grant No. 20100480839), Natu-
ral Science Foundation of Hebei Province (Grant No. A2010000184)
L. Guan et al. / Physics Letters A 375 (2011) 939945 945
and Natural Science Foundation of Hebei Province for Optoelec-
tronic Materials Research Base (Grant No. 08B007).
References
[1] D.M. Bagall, Y.F. Chen, Z. Zhu, T. Yao, S. Koyama, M.Y. Shen, T. Goto, Appl. Phys.
Lett. 70 (1997) 2230.
[2] D.-K. Hwang, S.-H. Kang, J.-H. Lim, E.-J. Yang, J.-Y. Oh, J.-H. Yang, S.-J. Park, Appl.
Phys. Lett. 86 (2005) 222101.
[3] W. Gao, Z.W. Li, J. Alloys Compd. 449 (2008) 202.
[4] M.H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo, P.
Yang, Science 292 (1897), (2001).
[5] B.S. Chua, S. Xu, Y.P. Ren, Q.J. Cheng, K. Ostrikov, J. Alloys Compd. 485 (2009)
379.
[6] S. Liang, H. Sheng, Y. Liu, Z. Huo, Y. Lu, H. Shen, J. Cryst. Growth 225 (2001)
110.
[7] T.V. Thu, S. Maenosono, J. Appl. Phys. 107 (2010) 014308.
[8] J.G. Lu, S. Fujita, T. Kawaharamura, H. Nishinaka, Y. Kamada, T. Ohshima, Z.Z.
Ye, Y.J. Zeng, Y.Z. Zhang, L.P. Zhu, H.P. He, B.H. Zhao, J. Appl. Phys. 101 (2007)
083705.
[9] X. Jiang, F.L. Wong, M.K. Fung, S.T. Lee, Appl. Phys. Lett. 83 (2003) 1875.
[10] S.D. Kirby, R.B. van Dover, Thin Solid Films 517 (2009) 1958.
[11] Y. Yang, J.J. Qi, Q.L. Liao, Y. Zhang, X.Q. Yan, Y.H. Huang, L.D. Tang, Appl. Phys.
A 94 (2009) 799.
[12] A.F. Kohan, G. Ceder, D. Morgan, Phys. Rev. B 61 (2000) 15019.
[13] J.G. Lu, Z.Z. Ye, F. Zhuge, Y.J. Zeng, B.H. Zhao, L.P. Zhu, Appl. Phys. Lett. 85 (2004)
3134.
[14] M. Kumar, T.H. Kim, S.S. Kim, B.T. Lee, Appl. Phys. Lett. 89 (2006) 112103.
[15] A. Allenic, X.Q. Pan, Y. Che, Z.D. Hu, B. Liu, Appl. Phys. Lett. 92 (2008) 022107.
[16] V. Vaithianathan, B.T. Lee, C.H. Chang, K. Asokan, S.S. Kim, Appl. Phys. Lett. 88
(2006) 112103.
[17] L.J. Mandalapu, Z. Yang, S. Chu, L. Liu, Appl. Phys. Lett. 92 (2008) 122101.
[18] D.C. Look, D.C. Reynolds, C.W. Litton, R.L. Jones, D.B. Eason, G. Cantwell, Appl.
Phys. Lett. 81 (2002) 1830.
[19] A.V. Singh, R.M. Mehra, J. Appl. Phys. 93 (2003) 396.
[20] T.M. Barnes, K. Olson, C.A. Wolden, Appl. Phys. Lett. 86 (2005) 112112.
[21] V. Vaithianathan, K. Asokan, J.Y. Park, S.S. Kim, Appl. Phys. A 94 (2009) 995.
[22] M.-S. Oh, D.-K. Hwang, Y.-S. Choi, J.-W. Kang, S.-J. Park, C.-S. Hwang, K.I. Cho,
Appl. Phys. Lett. 93 (2008) 111905.
[23] J.D. Ye, S.L. Gu, F. Li, S.M. Zhu, R. Zhang, Y. Shi, Y.D. Zheng, X.W. Sun, G.Q. Lo,
D.L. Kwong, Appl. Phys. Lett. 90 (2007) 152108.
[24] T.H. Vlasenin, M. Tanaka, Solid State Commun. 142 (2007) 292.
[25] F.X. Xiu, Z. Yang, L.J. Mandalapu, J.L. Liu, Appl. Phys. Lett. 88 (2006) 152116.
[26] K.-K. Kim, H.-S. Kim, D.-K. Hwang, J.-H. Lim, S.-J. Park, Appl. Phys. Lett. 83
(2003) 63.
[27] Z.G. Yu, P. Wu, H. Gong, Appl. Phys. Lett. 88 (2006) 132114.
[28] S. Limpijumnong, S.B. Zhang, S.-H. Wei, C.H. Park, Phys. Rev. Lett. 92 (2004)
155504.
[29] W.-J. Lee, J. Kang, K.J. Chang, Phys. Rev. B 73 (2006) 024117.
[30] M.D. McCluskey, S.J. Jokela, J. Appl. Phys. 106 (2009) 071101.
[31] A. Janotti, C.G. Van de Walle, Phys. Rev. B 76 (2007) 165202.
[32] C.H. Park, S.B. Zhang, S.-H. Wei, Phys. Rev. B 66 (2002) 073202.
[33] R.Y. Tian, Y.J. Zhao, J. Appl. Phys. 106 (2009) 043707.
[34] L. Guan, Q. Li, X. Li, J.X. Guo, B. Geng, L.T. Jin, Q.X. Zhao, J.W. Zhao, B.T. Liu, Adv.
Mater. Res. 7982 (2009) 1253.
[35] Z.G. Yu, H. Gong, P. Wu, Chem. Mater. 17 (2005) 852.
[36] H.V. Wenckstern, G. Benndorf, S. Heitsch, J. Sann, M. Brandt, H. Schmidt, J.
Lenzner, M. Lorenz, A.Y. Kuznetsov, B.K. Meyer, M. Grundmann, Appl. Phys.
A 88 (2007) 125.
[37] B.J. Kwon, H.S. Kwack, S.K. Lee, Y.H. Cho, D.K. Hwang, S.J. Park, Appl. Phys.
Lett. 91 (2007) 061903.
[38] M.D. Segall, P.J.D. Lindan, M.J. Probert, C.J. Pickard, P.J. Hasnip, J. Phys.: Condens.
Matter 14 (2002) 2717.
[39] J.P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 77 (1996) 3865.
[40] J.A. Dean, Langes Handbook of Chemistry, 14th ed., McGrawHill, New York,
1992.
[41] X.D. Zhang, M.L. Guo, W.X. Li, C.L. Liu, J. Appl. Phys. 103 (2008) 063721.
[42] Z.K. Tang, G.K.L. Wong, P. Yu, Appl. Phys. Lett. 72 (1998) 3270.
[43] G.E. Jellison Jr., L.A. Boatner, Phys. Rev. B 58 (1998) 3586.
[44] R.L. Hengehold, R.J. Almassy, Phys. Rev. B 1 (1970) 4784.
[45] D.J. Cai, J.Y. Kang, G.Y. Guo, Phys. Rev. B 80 (2009) 045311.
[46] Z.H. Levine, D.C. Allan, Phys. Rev. Lett. 66 (1991) 41.
[47] S.J. Clark, M.D. Segall, C.J. Pickard, P.J. Hasnip, M.J. Probert, K. Refson, M.C.
Payne, Zeitschrift fuer Kristallographie 220 (2005) 567.
[48] S.Zh. Karazhanov, P. Ravindran, A. Kjekshus, H. Fjellvg, B.G. Svensson, Phys.
Rev. B 75 (2007) 155104.
[49] J. Sun, H.T. Wang, J. He, Y. Tian, Phys. Rev. B 71 (2005) 125132.
[50] R.D. Shannon, Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystal-
logr. A 32 (1976) 751.
[51] P. Treguer, P. Le Corre, Deep-Sea Res. Part A 26 (1979) 163.
[52] Periodic Table of Elements, http://www.thelabrat.com/protocols/periodictable.
shtml.
[53] G.X. Hu, H. Gong, E.F. Chor, P. Wu, Appl. Phys. Lett. 89 (2006) 251102.
[54] D.W. Zeng, C.S. Xie, B.L. Zhu, W.L. Song, A.H. Wang, Mater. Sci. Eng. B 104
(2003) 68.
[55] M.D. Segall, R. Shah, C.J. Pickard, M.C. Payne, Phys. Rev. B 54 (1996) 16317.
[56] Y.W. Heo, K. Ip, S.J. Park, S.J. Pearton, D.P. Norton, Appl. Phys. A 78 (2004) 53.
[57] D.K. Hwang, M.S. Oh, J.H. Lim, C.G. Kang, S.J. Park, Appl. Phys. Lett. 90 (2007)
021106.
[58] H.T. Lin, T.S. Chin, J.C. Shih, S.H. Lin, T.M. Hong, R.T. Huang, F.R. Chen, J.J. Kai,
Appl. Phys. Lett. 85 (2004) 621.
[59] A. Schleife, C. Rdl, F. Fuchs, J. Furthmller, F. Bechstedt, Phys. Rev. B 80 (2009)
035112.
[60] H.Y. He, R. Orlando, M.A. Blanco, R. Pandey, Phys. Rev. B 74 (2006) 195123.