Valence-band electronic structure of CdO, ZnO, and MgO from x-ray photoemission spectroscopy and quasi-particle-corrected density-functional theory calculations
- SourceAvailable from: Tim Veal[show abstract] [hide abstract]
ABSTRACT: Metal oxides such as ZnO, Ga2O3, CdO, In2O3, and SnO2 exhibit high degree of transparency to visible light while supporting high levels electrical conductivity. The causes of the conductivity and the role played by the surface are current topics of research. This chapter presents a systematic study of the electronic structure and electrical properties of these post-transition metal oxides (PTMO) using a combination of X-ray photoelectron spectroscopy, angle-resolved photoelectron spectroscopy, Hall effect, infrared reflectivity, and optical absorption spectroscopy measurements. Evidence of surface electron accumulation in these PTMO is presented. It is found that for CdO and In2O3, electron accumulation is observed even in the absence of extremely high doping levels. The results also indicate that despite the strong tendency to exhibit surface electron accumulation, these materials can also exhibit an electron depletion layer under the appropriate surface stoichiometry conditions or when certain anions are adsorbed. The proclivity towards surface electron accumulation shown by the PTMOs is discussed in terms of bulk band structure, surface states, and the position of their band edges in an absolute energy scale. The electronic properties of thin films and bulk crystals of the PTMO surfaces also provide information vital for the interpretation of conductivity measurements of PTMO nanostructures, which are often dominated by surface effects.12/2011: pages 127-145;
Valence-band electronic structure of CdO, ZnO, and MgO from x-ray photoemission
spectroscopy and quasi-particle-corrected density-functional theory calculations
P. D. C. King,1,* T. D. Veal,1A. Schleife,2J. Zúñiga-Pérez,3B. Martel,3P. H. Jefferson,1F. Fuchs,2V. Muñoz-Sanjosé,4
F. Bechstedt,2and C. F. McConville1,†
1Department of Physics, University of Warwick, Coventry CV4 7AL, United Kingdom
2Institut für Festkörpertheorie und -Optik, Friedrich-Schiller-Universität, Max-Wien-Platz 1, D-07743 Jena, Germany
3Centre de Recherche sur l’Hétéro-Epitaxie et ses Applications, Centre National de la Recherche Scientifique, Parc de Sophia Antipolis,
Rue Bernard Grégory, 06560 Valbonne, France
4Departamento de Fisica Aplicada y Electromagnetismo, Universitat de Valéncia, C/Dr. Moliner 50, 46100 Burjassot, Spain
?Received 13 January 2009; revised manuscript received 23 February 2009; published 12 May 2009?
The valence-band density of states of single-crystalline rock-salt CdO?001?, wurtzite c-plane ZnO, and rock-
salt MgO?001? are investigated by high-resolution x-ray photoemission spectroscopy. A classic two-peak struc-
ture is observed in the VB-DOS due to the anion 2p-dominated valence bands. Good agreement is found
between the experimental results and quasi-particle-corrected density-functional theory calculations. Occupied
shallow semicore d levels are observed in CdO and ZnO. While these exhibit similar spectral features to the
calculations, they occur at slightly higher binding energies, determined as 8.8 eV and 7.3 eV below the valence
band maximum in CdO and ZnO, respectively. The implications of these on the electronic structure are
DOI: 10.1103/PhysRevB.79.205205PACS number?s?: 71.20.Nr, 79.60.?i, 71.15.Mb
Many II- and III-oxide materials have long found appli-
cation in polycrystalline form as transparent contacts in, for
example, photovoltaic devices, liquid-crystal displays, and
light-emitting diodes.1–3However, there is currently intense
interest in utilizing oxides as semiconductors in their own
optoelectronic4and high-performance electronic5device ap-
plications. While ZnO, with its room-temperature band gap
of ?3.3 eV and large exciton binding energy of ?60 meV,6
will likely form a central component of many II-O-based
optoelectronic devices, their spectral range can be extended
into the visible and deep ultraviolet by alloying ZnO with the
smaller band-gap compound CdO ?room-temperature Eg
?2.2 eV at the Brillouin-zone center7? and larger band-gap
compound MgO ?Eg?7.7 eV?.8,9Additionally, a high mo-
bility two-dimensional electron gas has already been demon-
strated in ZnO/MgZnO heterostructures,10indicating the po-
tential of this system for high-frequency electronic device
In order to fully exploit this materials system, it is neces-
sary to obtain a good understanding of the binary oxide com-
pounds, and in particular, their electronic structure. There
have been numerous theoretical investigations of electronic
band structure for II-O semiconductors and their alloys, uti-
lizing a variety of different calculation methods and correc-
tion schemes.11–19In order to test the validity of these calcu-
lations, however, it is necessary to compare the results to
experimentally measured quantities.
X-ray photoemission spectroscopy ?XPS? is a powerful
tool for investigating the electronic structure of solids. Pro-
vided a sufficiently large acceptance angle for detection of
emitted photoelectrons is used so that the whole Brillouin
zone is sampled, and given that final-state effects can gener-
ally be ignored, the photoemission intensity at low binding
energies gives the angle-integrated valence-band density of
states ?VB-DOS?,20–22weighted by the relative cross sections
for photoemission from the valence bands of given orbital
character. In many cases, however, at typical photon energies
for XPS measurements the cross sections for photoemission
from anion valence p states and cation valence s states that
commonly contribute to the topmost valence bands are ap-
proximately equal, and so the XPS spectra effectively probe
the total VB-DOS.22From the atomic photoionization cross-
section calculations of Yeh and Lindau,23this is also true for
the materials investigated in this work, and so it is a good
approximation to treat the photoemission spectra as giving
the total VB-DOS here. This is in contrast to studies of oxide
semiconductors by other spectroscopic techniques such as
soft x-ray emission ?SXE? which only yield the O 2p partial
density of states ?PDOS?. Additionally, XPS can easily be
used to investigate higher binding energy features such as
core levels and semicore levels.
Here, high-resolution XPS measurements are used to
probe the valence electronic structure of the II-O semicon-
ductors, rock-salt ?rs? CdO, wurtzite ?wz? ZnO, and rs-MgO.
In contrast to many of the previous investigations, in particu-
lar for CdO, single-crystalline thin-film samples are used,
allowing a direct determination of the electronic structure in
high-quality material of the form suitable for use in device
applications. The shallow semicore d levels in rs-CdO and
wz-ZnO, which are crucial in determining fundamental prop-
erties such as band gaps and valence-band offsets, are also
investigated. The XPS measurements show good agreement
with the results of first-principles calculations performed us-
ing density-functional theory ?DFT? incorporating quasipar-
II. EXPERIMENTAL AND THEORETICAL DETAILS
Single-crystalline CdO?001? was grown on r-plane sap-
phire to a thickness of ?500 nm by metal-organic vapor-
PHYSICAL REVIEW B 79, 205205 ?2009?
©2009 The American Physical Society205205-1
phase epitaxy at a growth temperature of ?380 °C, using
tertiary butanol and dimethylcadmium as the oxygen and
cadmium growth precursors, respectively. Single-crystalline
c-plane ZnO was grown on c-plane sapphire to a thickness of
?500 nm by plasma-assisted molecular-beam epitaxy at a
growth temperature of ?500 °C. Further details of the
elsewhere.24,25Single-crystalline bulk grown single-side pol-
ished MgO?001? was obtained from SPI Supplies.
High-resolution XPS measurements were performed us-
ing a Scienta ESCA300 spectrometer at the National Centre
for Electron Spectroscopy and Surface Analysis, Daresbury
Laboratory, U.K. X-rays of energy h?=1486.6 eV were pro-
duced using a monochromated rotating anode Al-K?x-ray
source. The ejected photoelectrons were analyzed by a 300
mm mean-radius spherical-sector electron energy analyzer
with 0.8 mm slits at a pass energy of 150 eV. The effective
instrumental resolution is 0.45 eV derived from the Gaussian
convolution of the analyzer broadening and the natural line-
width of the x-ray source ?0.27 eV?. For the insulating MgO,
charge compensation was achieved using a low-energy elec-
tron flood gun. In all cases, the binding-energy scale of the
photoemission measurements is referenced to the valence-
band maximum ?VBM?, determined by aligning the valence-
band photoemission with the broadened VB-DOS calcula-
tions, where the VBM is defined as 0 eV.
DFT calculations were performed using the hybrid func-
tional HSE03 for exchange and correlation.26The electron-
ion interaction was treated in the framework of the projector-
augmented wave method, including the shallow d electrons
in CdO and ZnO as valence states. Quasiparticle effects were
included in the calculation of the DOS by a G0W0correction
of the generalized Kohn-Sham eigenvalues. Details of the
calculation method and its application to the study of
II-O materials are reported elsewhere.16,27For comparison
with the experimental results, the quasi-particle-corrected
?QPC?-DFT DOS is broadened by a 0.2 eV full width at half
maximum ?FWHM? Lorentzian and a 0.45 eV FWHM
Gaussian to account for lifetime and instrumental broaden-
III. SURFACE PREPARATION
Core-level XPS spectra were recorded from the untreated
samples, and are shown for CdO in Fig. 1. A pronounced
multiple-peak structure was observed for the Cd 3d5/2core
level. The lower binding-energy component is attributed to
Cd-O bonding in the CdO, with the higher binding-energy
components attributed to bonding to more electronegative
species such as in CdO2?peroxide?, Cd?OH?2?hydroxide?,
and CdCO3?carbonate? compounds, present due to atmo-
spheric surface contamination and potentially remnants of
the growth precursors. Equivalently, the low binding-energy
component of the O 1s peak is attributed to Cd-O bonding,
with the higher binding-energy components due to the per-
oxide, hydroxide, and carbonate species. A large C 1s peak
was also observed, with a low binding-energy component
due to adventitious physisorbed hydrocarbon, and higher
binding-energy components due to carbonate and alcohol
species. The XPS core-level spectra for the ZnO and MgO
samples ?not shown? were very similar to those of the CdO.
To remove surface contamination, the samples were an-
nealed by electron-beam heating in a preparation chamber
connected to the XPS analysis chamber, at a temperature of
?600 °C for 2 h. Following annealing, the components in
the XPS core-level peaks due to peroxide, hydroxide and
carbonate species were quenched, with only a negligible
peak due to adventitious carbon remaining ?Fig. 1?. A slight
asymmetry to higher binding energies was observed on the
core-level peaks. However, CdO is known to exhibit electron
accumulation at its surface.28,29The asymmetry in the core-
level peaks is attributed to plasmon satellite features due to
conduction-band plasmons in the accumulation layer, as was
observed for the analogous compound, InN.30Similar results
were obtained for ZnO and MgO, although a slightly larger
peak due to adventitious carbon still remained following sur-
face preparation for the MgO sample.
IV. VALENCE-BAND ELECTRONIC STRUCTURE
Shirley-background-subtracted valence-band photoemis-
sion measurements from rs-CdO?001? are shown in Fig. 2?c?,
along with the calculated VB-DOS, with and without life-
time and instrumental broadening. The VB-DOS in this re-
gion arises from the three ?neglecting spin-orbit splitting?
topmost valence bands, the electronic structure of which is
shown in Fig. 2?d?.
The QPC-DFT DOS calculations contain much structure
in the valence-band region, due to the changing dispersions
and crossings of the electronic structure of the valence bands.
In particular, regions of energy where the bands are fairly flat
in k-space lead to peaks in the DOS, although the fine struc-
ture is largely smeared out by lifetime and instrumental
broadening. The resulting measured VB-DOS due to the top-
most three valence bands is characterized by a two-peak
structure, as observed from VB-XPS measurements of a
FIG. 1. ?Color online? Cd 3d5/2, O 1s, and C 1s XPS core-level
peaks from CdO before ?solid line? and after ?dashed line? anneal-
ing the sample at 600 °C in UHV. The intensities have been nor-
malized relative to the Cd 3d5/2intensity.
KING et al.
PHYSICAL REVIEW B 79, 205205 ?2009?
large number of other III–V and II–VI semiconductor com-
pounds by Ley et al.22This is despite the six fold coordina-
tion in this case, rather than the tetrahedral bonding for the
semiconductors investigated by Ley et al.22The agreement
of the valence photoemission with the broadened QPC-DFT
VB-DOS is good, in particular for the lower binding-energy
peak of the VB-DOS ?peak I?. The higher binding-energy
peak ?peak II? occurs at a slightly higher binding energy in
the experimental spectrum than in the calculation, and has a
lower intensity and larger width. By comparison with the
calculated valence-band structure, the main spectral weight
of peak II of the VB-DOS can be associated with turning
points around the bottom of the third valence band. This
suggests that the bottom of the third valence band may be
located slightly too shallow in energy by the QPC-DFT cal-
culations. Additionally, lifetime broadening is known to in-
crease with increasing binding energy,22although a constant
lifetime broadening has been applied to the QPC-DFT calcu-
lations here. This may explain the greater width and lower
intensity in peak II of the measured VB-DOS compared to
As discussed, the XPS measurements presented here ef-
fectively probe the total DOS of the system. It is of interest
to compare these results to the O 2p PDOS obtained previ-
ously from XES measurements on similar samples.31In par-
ticular, in the XES measurements, the ratio of peaks I to II is
approximately 2:1, whereas in the XPS measurements pre-
sented here this ratio is only ?1.3:1. This indicates that,
while peak I of the total VB-DOS may be dominated by
anion p-like contributions, peak II must contain significant
character from other orbitals. This is supported by the orbit-
ally resolved QPC-DFT calculations of the VB-DOS, shown
in Figs. 3?a? and 3?b?, which show, in particular, some Cd s-
and d-orbital character to peak II. This indicates that the third
valence band has appreciable cation s- and d-like character,
in contrast to the interpretation from simple tight-binding
A rather sharp onset of the VB-DOS occurs around the
VBM, which is broadened considerably by a combination of
lifetime and instrumental effects. Of particular interest here
is the nature of the calculated valence-band structure, which
indicates that the VBM does not occur at ?, but rather away
from ? at the L point, with another local maximum along the
? line between ? and K. These two critical points, located at
similar binding energy, give rise to the sharp onset in the
VB-DOS. CdO has occupied shallow d levels, as discussed
in detail in Sec. V. A pronounced p-d repulsion pushes the
valence-band states to higher energies. However, this p-d
repulsion is symmetry forbidden at ? for the octahedral point
symmetry of CdO’s rock-salt structure, causing the VBM to
occur away from ?, as discussed in detail elsewhere.13This
results in an indirect band gap, consistent with previous
theoretical11,13and experimental32,33investigations. The hy-
bridization of the p and d orbitals is also evident from the
Cd d-orbital character present in the VB-DOS close to the
VBM, as shown in Fig. 3.
The magnitude of the indirect band gap has been studied
extensively experimentally, although there has been a large
spread in the results obtained. Values of 0.55 ?Ref. 32? and
0.84 eV ?Ref. 33? are widely quoted34for the indirect band
gap at room temperature and at 100 K, respectively. How-
ever, McGuinness et al.,35interpreting the previous photo-
emission results of Dou et al.,36,37suggested a value of
?1.2 eV, while their own x-ray emission and absorption
measurements suggested a value of almost 2 eV, although
this determination is complicated by limited resolution and
the elementally specific nature of the measurement. The
FIG. 2. ?Color online? Shirley-background-subtracted photo-
emission spectra from ?a? around the Cd 4d peaks, and ?c? the va-
lence band, and QPC-DFT VB-DOS calculations shown without
?shaded? and with lifetime and instrumental broadening for rs-CdO.
The corresponding QPC-DFT band structure for rs-CdO is shown in
?b? and ?d?. The photoemission spectra and broadened QPC-DFT
calculations are normalized to the same maximum value both for ?a?
the d-level region and ?c? the valence-band region. The main fea-
tures in the VB-DOS are marked after Ley et al. ?Ref. 22?.
FIG. 3. ?Color online? Total and s-, p-, and d-resolved cation-
and anion-projected VB-DOS for ??a?,?b?? CdO, ??c?,?d?? ZnO, and
??e?,?f?? MgO from QPC-DFT calculations.
VALENCE-BAND ELECTRONIC STRUCTURE OF CdO,…
PHYSICAL REVIEW B 79, 205205 ?2009?
good agreement of the spectral features in the calculated VB-
DOS and the valence-band photoemission, particularly
around the lower binding-energy peak, suggests that the en-
ergies of critical points in the valence-band electronic struc-
ture are accurately determined by the QPC-DFT calculations
presented here. These calculations give the L3
?the energy difference of the top valence band at the VBM
?L-point? and at the zone center? as 1.09 eV. Taking the
room-temperature band gap as 2.16 eV,7this gives an indi-
rect band gap at room temperature of 1.07 eV, significantly
higher than the widely quoted value of 0.55 eV.
A third peak was also observed in the valence-
photoemission measurements of Ley et al.,22associated with
the fourth, largely anion s-like, valence band. However, this
is not observed here within the binding-energy range inves-
tigated. This is due to the highly ionic nature of CdO, which
leads to a large ionicity gap, with the fourth valence band
located a long way below the VBM ??18 eV in the theoret-
ical calculations?. Due to its highly bound nature, it takes on
a somewhat semicore-level character of an O 2s orbital, with
only limited dispersion of the band throughout the Brillouin
zone, representing its localized nature in real space. This
band is not considered further in the work presented here.
The stable polymorph of ZnO is the wurtzite, rather than
the rock-salt crystal structure. This leads to pronounced dif-
ferences in the valence-band electronic structure of ZnO,
shown in Fig. 4?d?, as compared to CdO ?Fig. 2?d??. In par-
ticular, the presence of a nonzero ?albeit small? crystal-field
splits the top three valence bands into six bands. Addition-
ally, although ZnO also has occupied d levels, p-d repulsion
is now symmetry allowed at ?, resulting in a single VBM
Despite these differences in the electronic band structure,
the calculated VB-DOS ?Fig. 4?b?? shows rather similar fea-
tures to that of CdO, with two main peaks now due largely to
the uppermost four and next two valence bands, respectively.
The onset of the VB-DOS is rather more gradual around the
VBM than for CdO. This is because there is only one critical
point in the band structure around the VBM in ZnO, whereas
the VBM and a local maximum at similar binding energies
both contributed to the onset of the VB-DOS in CdO.
The valence photoemission spectrum is compared to the
broadened VB-DOS calculations in Fig. 4?c?. Excellent
agreement is obtained between experiment and theory, indi-
cating that the QPC-DFT calculations accurately determine
the VB-DOS, and by extension the correct valence-band
structure for ZnO. In this case, even for peak II of the VB-
DOS, there is good agreement between experiment and
theory, both in terms of binding energy and relative intensity
and width compared to peak I.
As for CdO, it is of interest to compare the XPS results
determined here with previous SXE measurements of the
O 2p PDOS.38Again, the intensity ratio of peaks I and II is
much larger in the SXE measurements than in the XPS mea-
surements, indicating an appreciable component of peak II of
the VB-DOS that is not of anion p-like character. This is
supported by the orbitally resolved VB-DOS calculations,
shown in Figs. 3?c? and 3?d?, indicating that peak II of the
total VB-DOS, and hence the corresponding valence bands,
has considerable anion s- and d-orbital character.
As for CdO, the stable polymorph of MgO is the rock-salt
structure. However, Mg has no occupied d orbitals and so, as
shown in Fig. 5?b?, the VBM occurs at ?. The VB-DOS due
to the top three valence bands ?there is no crystal-field split-
ting for the rock-salt structure?, shown in Fig. 5?a?, contains
the same two-peak structure seen above for CdO and ZnO,
although there is now a pronounced shoulder on the low
binding-energy side of peak I due to critical points in the
valence-band structure around K, W, and between L and X.
The general features of the calculated VB-DOS are well re-
produced by experiment, in particular the binding energies of
the peaks. However, the XPS yields a somewhat broader
spectrum than the broadened QPC-DFT VB-DOS. This may
be due to a larger lifetime broadening for MgO than for the
other II-O materials considered above, but may also be an
experimental artifact. As discussed above, some carbon con-
tamination still remained on the MgO sample after surface
preparation, and this may have led to a slight broadening of
the spectral features. Also, although charge compensation
was performed using a low-energy electron flood gun, a
slight broadening of the experimental spectrum may result if
a small amount of charging remained.
Comparing the XPS measurements presented here with
previous SXE measurements,27it is again clear that peak II
has some intensity that cannot be attributed to valence bands
of anion 2p character, and the calculations shown in
FIG. 4. ?Color online? Shirley-background-subtracted photo-
emission spectra from ?a? around the Zn 3d peaks, and ?c? the va-
lence band, and QPC-DFT VB-DOS calculations shown without
?shaded? and with lifetime and instrumental broadening for wz-
ZnO. The corresponding QPC-DFT band structure for wz-ZnO is
shown in ?b? and ?d?. The photoemission spectra and broadened
QPC-DFT calculations are normalized to the same maximum value
both for ?a? the d-level region and ?c? the valence-band region. The
main features in the VB-DOS are marked after Ley et al. ?Ref. 22?.
KING et al.
PHYSICAL REVIEW B 79, 205205 ?2009?
Figs. 3?e? and 3?f? indicate some cation s- and p-orbital char-
acter in this region. However, there are now no d-orbitals
which can contribute to this peak, as for CdO and ZnO.
V. SEMICORE d-LEVELS
As discussed above, the presence ?CdO and ZnO? or ab-
sence ?MgO? of occupied d-orbitals shallow in binding en-
ergy can have a profound influence on the valence-band elec-
tronic structure. In particular, their influence on the VBM
position can have significant consequences for the band
offsets18,19and also band gaps19,39of these materials. It is
therefore of interest to consider these levels in some detail.
Photoemission measurement from around the Cd 4d
?Zn 3d? peaks of CdO ?ZnO? are shown in Fig. 2?a? ?Fig.
4?a??, compared to QPC-DFT calculations of the DOS, and
the corresponding calculated electronic structure ?Fig. 2?c?
and Fig. 4?c??. The band structure is somewhat more compli-
cated in this region for ZnO than CdO due to the crystal-field
splitting. Despite their low binding energies, the shallow dis-
persion of the bands indicate their localized semicore-level
character, which leads to intense, narrow peaks in the DOS.
These are significantly reduced in intensity and increased in
width by lifetime and instrumental broadening effects. How-
ever, in both cases, the width of these features is much larger
in the photoemission measurements than in the theory. This
is likely due to an increase in lifetime broadening for these
semicore-level features as compared to the much less local-
ized valence states, which has not been included in the
broadening of the theoretical calculations. There are also
small high binding-energy shoulders on these levels, attrib-
uted to energy losses to conduction-band plasmons in
electron-accumulation layers, as was previously observed for
the analogous material, InN.30,40
Additionally, it is evident that the binding energy of the
Cd 4d-like and Zn 3d-like semicore levels are slightly under-
estimated by the theoretical calculations compared to the ex-
perimental measurements. This slight underestimation of
shallow d-level positions has previously been observed for
QPC-DFT calculations utilizing the HSE03 functional,16,40
and these results suggest that it may be a rather general fea-
ture of this calculation scheme.
It is useful to determine the position of these levels ex-
perimentally. However, how this should be achieved is not
well defined. The binding energy of core-level features in
XPS measurements is usually determined by peak-fitting
Voigt ?mixed Lorentzian-Gaussian? line shapes to the spec-
tral features, properly accounting for effects such as stepped
backgrounds, and spin-orbit split components.41However, it
is not appropriate to treat the semicore d levels in this man-
ner. This can be effectively illustrated by considering the
Cd 4d-like levels. If these were assumed to be true core lev-
els, they should be described by a spin-orbit split doublet,
with the intensity ratio of the lower binding energy ?4d5/2? to
higher binding energy ?4d3/2? components being 3:2. How-
ever, from Fig. 2?a?, it is clear that this intensity ratio is not
observed, indicating that these cannot be considered as con-
ventional core levels, and standard spectral functions cannot
be used in their analysis. The position of these levels is there-
fore defined here as the binding energy of the maximum
intensity of these features. Consequently, the binding energy
of the semicore Cd 4d levels in rs-CdO is determined as 8.8
eV below the VBM, while the Zn 3d levels in wz-ZnO occur
at 7.3 eV below the VBM, in contrast to values of 8.3 and 6.9
eV determined from the QPC-DFT calculations. The experi-
mental values determined here are in reasonable agreement
with the position of the O 2p-Cd 4d and O 2p-Zn 3d hybrid-
ized components observed in SXE experiments.31,38
The valence-band electronic structure of the II-VI oxides,
CdO, ZnO, and MgO, has been investigated using a combi-
nation of high-resolution x-ray photoemission spectroscopy
functional theory calculations. The valence-band density of
states of each compound can broadly be characterized by a
two-peak structure, dominated by anion 2p contributions, al-
though with contributions from other cation orbitals, particu-
larly in the higher binding-energy peak. There are also quali-
tative differences between the different compounds due
largely to variations in crystal structures, and the presence or
absence of occupied cation d orbitals. Good agreement be-
tween the experimental and theoretical results was obtained
in all cases. The occupied shallow semicore d levels were
also investigated in CdO and ZnO, and their binding energies
were found to be 8.8 eV and 7.3 eV below the valence-band
maximum, respectively. These values were slightly underes-
valence-band photoemission spectrum, and QPC-DFT VB-DOS
calculations shown without ?shaded? and with lifetime and instru-
mental broadening for rs-MgO. The corresponding QPC-DFT
valence-band structure for rs-MgO is shown in ?b?. The photoemis-
sion spectra and broadened QPC-DFT calculations are normalized
to the same maximum value. The main features in the VB-DOS are
marked after Ley et al. ?Ref. 22?.
?Color online? ?a? Shirley-background-subtracted
VALENCE-BAND ELECTRONIC STRUCTURE OF CdO,…
PHYSICAL REVIEW B 79, 205205 ?2009?
timated in the theoretical calculations in comparison to the
experiment, although the calculated spectral shape of these
features was similar to that found experimentally.
We are grateful to D. Law and G. Beamson of NCESS for
their assistance with XPS measurements. Also, we acknowl-
edge the Engineering and Physical Sciences Research Coun-
cil, U.K., for financial support under Grant No. EP/
E031595/1 and access to the NCESS facility under Grant
No. EP/E025722/1, the Spanish Government for financial
support under Grant No. MAT2007-66129, the Deutsche
Forschungsgemeinschaft for financial support under Project
No. Be1346/20-1, and the Carl-Zeiss-Stiftung.
1I. Hamberg and C. G. Granqvist, J.Appl. Phys. 60, R123 ?1986?.
2T. J. Coutts, D. L. Young, X. Li, W. P. Mulligan, and X. Wu, J.
Vac. Sci. Technol. A 18, 2646 ?2000?.
3T. Minami, Semicond. Sci. Technol. 20, S35 ?2005?.
4A. Tsukazaki et al., Nature Mater. 4, 42 ?2005?.
5A. P. Ramirez, Science 315, 1377 ?2007?.
6Ü. Özgür, Y. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S.
Doğan, V. Avrutin, S.-J. Cho, and H. Morkoç, J. Appl. Phys. 98,
7P. H. Jefferson, S. A. Hatfield, T. D. Veal, P. D. C. King, C. F.
McConville, J. Zúñiga-Pérez, and V. Muñoz-Sanjosé, Appl.
Phys. Lett. 92, 022101 ?2008?.
8T. Makino, Y. Segawa, M. Kawasaki, A. Ohtomo, R. Shiroki, K.
Tamura, T. Yasuda, and H. Koinuma, Appl. Phys. Lett. 78, 1237
9S. Sadofev, S. Blumstengel, J. Cui, J. Puls, S. Rogaschewski, P.
Schäfer, and F. Henneberger, Appl. Phys. Lett. 89, 201907
10A. Tsukazaki, A. Ohtomo, T. Kita, Y. Ohno, H. Ohno, and M.
Kawasaki, Science 315, 1388 ?2007?.
11J. E. Jaffe, R. Pandey, and A. B. Kunz, Phys. Rev. B 43, 14030
12Y. Dou, R. G. Egdell, D. S. L. Law, N. M. Harrison, and B. G.
Searle, J. Phys.: Condens. Matter 10, 8447 ?1998?.
13A. Schleife, F. Fuchs, J. Furthmuller, and F. Bechstedt, Phys.
Rev. B 73, 245212 ?2006?.
14A. Janotti, D. Segev, and C. G. Van de Walle, Phys. Rev. B 74,
15D. Fritsch, H. Schmidt, and M. Grundmann, Appl. Phys. Lett.
88, 134104 ?2006?.
16F. Fuchs, J. Furthmüller, F. Bechstedt, M. Shishkin, and G.
Kresse, Phys. Rev. B 76, 115109 ?2007?.
17S. Z. Karazhanov, P. Ravindran, A. Kjekshus, H. Fjellvag, and
B. G. Svensson, Phys. Rev. B 75, 155104 ?2007?.
18A. Janotti and C. G. Van de Walle, Phys. Rev. B 75, 121201?R?
19Y. Z. Zhu, G. D. Chen, H. Ye, A. Walsh, C. Y. Moon, and S.-H.
Wei, Phys. Rev. B 77, 245209 ?2008?.
20D. A. Shirley, Phys. Rev. B 5, 4709 ?1972?.
21R. A. Pollak, L. Ley, S. Kowalczyk, D. A. Shirley, J. D. Joan-
nopoulos, D. J. Chadi, and M. L. Cohen, Phys. Rev. Lett. 29,
22L. Ley, R. A. Pollak, F. R. McFeely, S. Kowalczyk, and D. A.
Shirley, Phys. Rev. B 9, 600 ?1974?.
23J. J. Yeh and I. Lindau, At. Data Nucl. Data Tables 32, 1 ?1985?.
24J. Zúñiga-Pérez, C. Munuera, C. Ocal, and V. Muñoz-Sanjosé, J.
Cryst. Growth 271, 223 ?2004?.
25F. Vigué, P. Vennéguès, C. Deparis, S. Vézian, M. Laügt, and
J.-P. Faurie, J. Appl. Phys. 90, 5115 ?2001?.
26J. Heyd, G. E. Scuseria, and M. Ernzerhof, J. Chem. Phys. 118,
27A. Schleife, C. Rödl, F. Fuchs, J. Furthmüller, F. Bechstedt, P. H.
Jefferson, T. D. Veal, C. F. McConville, L. F. J. Piper, A. De-
Masi, K. E. Smtih, H. Lösch, R. Goldhahn, C. Cobet, J. Zúñiga-
Pérez, and V. Muñoz-Sanjosé, J. Korean Phys. Soc. 53, 2811
28L. F. J. Piper, L. Colakerol, P. D. C. King, A. Schleife, J. Zúñiga-
Pérez, P. A. Glans, T. Learmonth, A. Federov, T. D. Veal, F.
Fuchs, V. Muñoz-Sanjosé, F. Bechstedt, C. F. McConville, and
K. E. Smith, Phys. Rev. B 78, 165127 ?2008?.
29P. D. C. King, T. D. Veal, P. H. Jefferson, J. Zúñiga-Pérez, V.
Muñoz-Sanjosé, and C. F. McConville, Phys. Rev. B 79, 035203
30P. D. C. King, T. D. Veal, H. Lu, S. A. Hatfield, W. J. Schaff, and
C. F. McConville, Surf. Sci. 602, 871 ?2008?.
31L. F. J. Piper, A. DeMasi, K. E. Smith, A. Schleife, F. Fuchs, F.
Bechstedt, J. Zuniga-Pérez, and V. Munoz-Sanjosé, Phys. Rev. B
77, 125204 ?2008?.
32H. Köhler, Solid State Commun. 11, 1687 ?1972?.
33F. P. Koffyberg, Phys. Rev. B 13, 4470 ?1976?.
34H. Finkenrath, in Physics of II-VI and I-VII Compounds, Semi-
Magnetic Semiconductors, Landolt-Börnstein, New Series,
Group III: Crystal and Solid State Physics, Vol. 17B, edited by
O. Madelung, M. Schulz, and H. Weiss ?Springer, Berlin, 1982?.
35C. McGuinness, C. B. Stagarescu, P. J. Ryan, J. E. Downes, D.
Fu, K. E. Smith, and R. G. Egdell, Phys. Rev. B 68, 165104
36Y. Dou, T. Fishlock, R. G. Egdell, D. S. L. Law, and G. Beam-
son, Phys. Rev. B 55, R13381 ?1997?.
37Y. Dou, R. G. Egdell, T. Walker, D. S. L. Law, and G. Beamson,
Surf. Sci. 398, 241 ?1998?.
38A. R. H. Preston, B. J. Ruck, L. F. J. Piper, A. DeMasi, K. E.
Smith, A. Schleife, F. Fuchs, F. Bechstedt, J. Chai, and S. M.
Durbin, Phys. Rev. B 78, 155114 ?2008?.
39P. Carrier and S.-H. Wei, J. Appl. Phys. 97, 033707 ?2005?.
40P. D. C. King, T. D. Veal, C. F. McConville, F. Fuchs, J. Furth-
müller, F. Bechstedt, J. Schörmann, D. J. As, K. Lischka, H. Lu,
and W. J. Schaff, Phys. Rev. B 77, 115213 ?2008?.
41D. Briggs and J. C. Rivière, in Auger and X-ray Photoelectron
Spectroscopy, Practical Surface Analysis, 2nd ed., Vol. 1 edited
by D. Briggs and M. P. Seah ?Wiley, Chichester, 1990?,
KING et al.
PHYSICAL REVIEW B 79, 205205 ?2009?