Electronic structure of ReS2 and ReSe2 from first-principles calculations, photoelectron spectroscopy, and electrolyte electroreflectance

Article abstract of DOI:10.1103/PhysRevB.60.15766

The electronic structures of ReS₂ and ReSe₂ single crystals are investigated using a first-principles density-of-states (DOS) calculation, ultraviolet photoelectron spectroscopy (UPS), and electrolyte electroreflectance (EER). The total and partial DOS were calculated by the full-potential linearized-augmented-plane-wave method. From the calculations, the main contribution near the band edge of Re X₂ (X=S,Se) is determined to be dominated by the nonbonding Re d orbitals. The valence-band DOS is experimentally verified by the UPS measurements. EER measurements were performed in the energy range of 1.3-6 eV. The EER spectra exhibit sharp derivativelike structures in the vicinity of the band-edge excitonic transitions as well as higher-lying interband transitions. Transition energies are determined accurately. From the experimental and the theoretically calculated results, probable energy-band structures of ReS₂ and ReSe₂ are constructed. 1999 The American Physical Society.

Full text of DOI:10.1103/PhysRevB.60.15766

PHYSICAL REVIEW B
VOLUME 60, NUMBER 23
15 DECEMBER 1999-I
Electronic structure of ReS2 and ReSe2 from first-principles calculations,
photoelectron spectroscopy, and electrolyte electroreflectance
†
*
C. H. Ho and Y. S. Huang
Department of Electronic Engineering, National Taiwan University of Science and Technology, Taipei 106, Taiwan, Republic of China
J. L. Chen and T. E. Dann
Synchrotron Radiation Research Center, Hsinchu 300, Taiwan, Republic of China
K. K. Tiong
Department of Electrical Engineering, National Taiwan Ocean University, Keelung 202, Taiwan, Republic of China
͑Received 27 May 1999͒
The electronic structures of ReS2 and ReSe2 single crystals are investigated using a first-principles density-
of-states ͑DOS͒ calculation, ultraviolet photoelectron spectroscopy ͑UPS͒, and electrolyte electroreflectance
͑EER͒. The total and partial DOS were calculated by the full-potential linearized-augmented-plane-wave
method. From the calculations, the main contribution near the band edge of Re X2 (XϭS,Se) is determined to
be dominated by the nonbonding Re d orbitals. The valence-band DOS is experimentally verified by the UPS
measurements. EER measurements were performed in the energy range of 1.3–6 eV. The EER spectra exhibit
sharp derivativelike structures in the vicinity of the band-edge excitonic transitions as well as higher-lying
interband transitions. Transition energies are determined accurately. From the experimental and the theoreti-
cally calculated results, probable energy-band structures of ReS2 and ReSe2 are constructed.
͓S0163-1829͑99͒07847-9͔
was first determined by Alcock and Kjekshus.9 They pro-
I. INTRODUCTION
posed that the distortion of ReSe2 bonds was related to the
ReS2 and ReSe2 are diamagnetic indirect semiconductors
belonging to the family of transition-metal dichalcogenides
crystallized in a distorted layered structure of triclinic sym-
movement of the rhenium atoms toward each other. Wilder-
vanck and Jellinek1 showed that ReS2 and ReSe2 are isos-
tructural and crystallized in a distorted CdCl2 structure of
triclinic symmetry. Lamfers et al.10 reported the structural
data of Re X2 using single-crystal x-ray diffraction and
claimed that for ReS2 there are two sandwiches in a unit cell
related by symmetry center. However, from the recent study
of the crystal structure of ReS2ϪxSex layered series by pow-
der x-ray diffraction,11 the whole series of ReS2ϪxSex single
crystals were shown to be single phase and isostructural. The
structural data of binary compounds ReS2 and ReSe2 were
determined to be similar to the results of Wildervanck and
Jellinek.1
The energy-band structure is one of the most important
factors in determining the solid-state properties of a material.
In this paper we evaluate the electronic structures of ReS2
and ReSe2 using the density-of-states ͑DOS͒ calculations, the
ultraviolet photoelectron spectroscopy ͑UPS͒, and
electrolyte-electroreflectance ͑EER͒ measurements. The total
and partial density of states of the semiconducting layered
compounds ReS2 and ReSe2 were numerically calculated
self-consistently with the full-potential linearized-
augmented-plane-wave ͑LAPW͒ method.12 For the experi-
mental analysis, UPS is employed to extract information on
the valence-band electronic structure of the material. The
electronic band structure of the material is being probed by
the EER technique. The EER technique has been proven to
be a very powerful tool in the study of the band structure of
semiconductors.13,14 The derivative nature of EER spectra
suppresses uninteresting background effects and greatly en-
¯
metry ͑space group P1͒. In recent investigations, these com-
pounds have been attractive as electrode or photoelectrode
materials because of their catalytic properties and favorable
stability.1–4 Photoelectrochemical spectral response measure-
ments showed that ReS2 and ReSe2 single crystals had an
enhanced stability against photocorrosion with respect to
other semiconductors such as CdS and GaAs.5 The chemical
stability is mainly attributed to the chemically saturated sur-
face of the crystals in connection with some d character of
the bonding. Furthermore, ReS2 is of considerable interest as
a catalytic material having potential application as a sulfur-
tolerant hydrogenation and hydrodesulfurization catalyst1,6
and as a promising solar-cell material in electrochemical
cells.4,7
Re X2 can be thought of as distorted 1T-MX2
dichalcogenides.6 The 1T-MX2 phase consists of edge-
shared MX6 octahedra. In each MX2 layer, the metal-atom
sheet is sandwiched by chalcogen-atom sheets and the metal
atoms of an undistorted MX2 layer form a hexagonal lattice.
Re X2 have a d3 electron count; the metal atoms in the metal
sheet slip off their regular sites comprising a Re4 ‘‘diamond
unit’’ and resulting in a shorter Re-Re distance.8 These Re4
‘‘diamonds’’ are coplanar and coupled with one another to
form a one-dimensional clustering pattern of ‘‘diamond
chains.’’ The diamond-chain clusters in the metal sheets of
Re X2 resulted in a lattice distortion from the perfect octahe-
dral layered structure. The structural information on ReSe2
0163-1829/99/60͑23͒/15766͑6͒/$15.00
PRB 60
15 766
©1999 The American Physical Society

PRB 60
15 767
TABLE I. Input parameters for the DOS calculations of Re X2 .
ELECTRONIC STRUCTURE OF ReS2 AND . . .
hances the precision in the determination of transition ener-
gies. The shaper line shapes as compared to the conventional
optical techniques have enabled us to achieve a greater reso-
lution and hence to detect weaker features. Subsequently,
from the EER spectra of ReS2 and ReSe2 we observed more
features which were not detected by the UPS measurements.
The EER spectra can be fitted with a form of the Aspnes
equation of the derivative Lorentzian line shape.15 From a
detailed line-shape fit, the transition energies of the band-
edge excitonic and higher-lying interband transitions are de-
termined accurately. From the experimentally and theoreti-
cally calculated results, together with the results of previous
optical-absorption measurements,16 probable band-structure
schemes for ReS2 and ReSe2 are constructed.
Material
ReS2
ReSe2
Lattice parameters
a ͑Å͒
b ͑Å͒
c ͑Å͒
␣ ͑deg͒
␤ ͑deg͒
␥ ͑deg͒
6.450
6.390
6.403
105.49
91.32
119.03
6.713
6.623
6.740
104.59
92.28
118.79
¯
P1
¯
P1
Space group
information on the density of states in the electronic structure
of materials. In this study, the ultraviolet light source was
derived from the synchrotron radiation source by the LSGM
beamline at the Synchrotron Radiation Research Center
͑SRRC͒. The light source provides an ultraviolet ray in the
energy range of 15–200 eV and a spot size of ϳ1.5
ϫ1.5 mm2. The crystals were cleaved and placed in a highly
evacuated chamber with a pressure of ϳ8ϫ10Ϫ11 Torr. The
monochromatic ultraviolet beam is filtered by a spherical
grating monochromator, and a VSW hemispherical collector
of a multichannel analyzer collected the emitted photoelec-
trons. The measurements were done on an as-grown ͑001͒
surface with the photon incidence angle kept at 20° and the
detection angle at 45°. The incident photon energies for mea-
suring the photoelectrons of the valence band and Re 4f core
levels were fixed at 50 and 100 eV, respectively. The UPS
spectra were deduced from the electron counts in various
channels of the analyzer with an energy resolution of
0.05 eV.
The EER spectra were taken on a fully computerized
setup for modulation spectroscopy described elsewhere.17
The detector response to the dc component of the reflected
light is kept constant by either an electronic servo mecha-
nism or a neutral density filter so that the ac reflectance cor-
responds to ⌬R/R, the differential reflectance. Scans of
⌬R/R versus wavelength are obtained using a 0.35 m
McPherson grating monochromator together with an Oriel
150 W xenon arc lamp as a monochromatic light source.
Phase-sensitive detection is used to measure the differential
reflectance. Plate-shaped crystals were selected for EER
measurements. The electrolyte was a 1 N H2SO4 aqueous
solution, and the counter electrode was a 5 cm2 platinum ͑Pt͒
plate. A 200 Hz, 100 mV peak-to-peak square wave with
Vdcϭ0 V versus Pt electrode was used to modulate the elec-
tric field in the space-charge region of the ReS2 or ReSe2
electrodes. The magnitude of the modulated field across the
space-charge region must be maintained such that the EER
line shape remains invariant and the amplitude of ⌬R/R var-
ies linearly with the modulation voltages.
II. EXPERIMENTAL DETAIL AND DOS CALCULATION
Single crystals of ReS2 and ReSe2 were grown using the
chemical vapor transport method with Br2 as the transport
agent. Prior to crystal growth, quartz tubes containing bro-
mine and the elements ͑Re: 99.95% pure, S: 99.999%,
Se: 99.999%͒ were evacuated and sealed. To improve the
stoichiometry, sulfur or selenium with 2 mol % in excess was
added with respect to rhenium. The quartz tube was placed in
a three-zone furnace and the charge prereacted for 24 h at
800 °C while the temperature of the growth zone was set at
1000 °C to prevent the transport of the product. The furnace
was then equilibrated to give a constant temperature across
the reaction tube, and was programmed over 24 h to produce
the temperature gradient at which single-crystal growth takes
place. The best results were obtained with temperature gra-
dients of about 1060→1010 °C for ReS2 and 1050
→1000 °C for ReSe2. Both ReS2 and ReSe2 formed thin,
silver-colored, graphitelike, hexagonal platelets up to 2 cm2
in area and 100 ␮m in thickness. X-ray-diffraction patterns
of single crystals were obtained using Ni-filtered Cu K␣ ra-
diation. The patterns confirmed the triclinic symmetry of
ReS2 and ReSe2 with all parameters consistent with those
previously reported.2,9 Electron probe microanalysis indi-
cated a chalcogen deficiency in the crystals. Hall effect mea-
surements revealed n-type semiconducting behavior.
Optical-absorption measurements showed indirect semicon-
ducting behavior with an energy gap of 1.37 eV for ReS2 and
1.19 eV for ReSe2.
For the computational work, we utilized WIEN97 software
to calculate the electronic band structure and density of states
for ReS2 and ReSe2. This program package is capable of
performing the electronic structure calculations of solids us-
ing the full-potential LAPW method.12 Since the layered ma-
terials are characterized by strong covalent intralayer bond-
ing and weak van der Waals interlayer interactions, the
calculation of the electronic band structure for single-layer
Re X2 was employed. In this paper the electronic-structure
calculations were performed using LAPW method with the
structural data from Table I as the input parameters. By in-
corporating the information of lattice parameters, space
group, and atomic coordination of ReS2 and ReSe2 into the
WIEN97 software, the partial density of states of Re atoms and
chalcogen atoms and total density of states of Re X2 com-
pounds are respectively determined.
III. RESULTS AND DISCUSSION
Displayed in Figs. 1͑a͒ and 1͑b͒ are the UPS spectra and
the calculated DOS for ReS2 and ReSe2 in the energy range
near the valence band. The dotted lines in Fig. 1 correspond
to the DOS of the compounds Re X2 ͑XϭS or Se͒, the solid
lines are those of metal Re atoms, and the dashed lines are
the calculated DOS of chalcogen atoms. The experimental
The energy distribution of the photoelectrons will provide

15 768
HO, HUANG, CHEN, DANN, AND TIONG
PRB 60
FIG. 2. Ultraviolet photoelectron spectra of the Re 4f core levels
of ReS2 and ReSe2.
Re 5d eg /X p complexes. From the DOS calculations, the en-
ergy gaps of ReS2 and ReSe2 are determined to be 1.16 and
0.89 eV, respectively. The gap energy of 0.89 eV for ReSe2
is in approximate agreement with the value of 0.87 eV cal-
culated for the same crystal by Kertesz et al.8 where a simple
¨
tight-binding band-structure calculation of extended Huckel
type is being employed. Recently, Raybaud et al.18 reported
ab initio density-functional studies of the electronic structure
FIG. 1. The experimental UPS spectra ͑upper parts͒ and calcu-
lated density of states ͑lower parts͒ for ͑a͒ ReS2 and ͑b͒ ReSe2. The
dotted lines in the lower parts correspond to the DOS of the com-
pounds Re X2 ͑XϭS or Se͒, the solid lines are those of metal Re
atoms, and the dashed lines are the calculated DOS of chalcogen
atoms.
UPS spectra and the calculated DOS exhibit a multitude of
sharp features which are due to the triclinic low-symmetry
structures of Re X2. Figure 1 also indicates a broad and flat
region of the valence-band UPS spectra of Re X2, which is
due to the three d electrons of the rhenium atom whose effect
is extended over the entire valence band and strongly hybrid-
ized with the S 3p (Se 4p) states. This is a consequence of
a stronger overlap of the wave functions of Re 5d orbitals
owing to the shorter Re-Re distances. Through UPS mea-
surements, the work functions of the materials can be deter-
mined from the minimum photoionization energies of the
collected photoelectrons. The work functions for ReS2 and
ReSe2 are estimated to be 5.75 and 5.6 eV, respectively.
From the DOS calculation, the widths of the valence band
are 7.05 eV for ReS2 and 6.9 eV for ReSe2. These results
agreed well with the UPS measurements. For Re X2, the S 3s
(Se 4s) band ranges from Ϫ13.2 to Ϫ16.1 eV ͑Ϫ13.4 to
Ϫ16.3 eV͒ with respect to the Fermi level located at the
bottom of the conduction band. The calculated results also
reveal that the contributions of Re atoms ͑solid lines͒ near
the top portions of the valence band and the lowest portions
of the conduction band are much larger than those of chal-
cogen atoms ͑dashed lines͒. The other parts of the valence
and conduction bands are dominated by the contributions of
FIG. 3. The EER spectra ͑dashed lines͒ of ͑a͒ ReS2 and ͑b͒
ReSe2 at room temperature in the range of 1.3–6.0 eV. The solid
lines are least-squares fits to the Aspnes derivative line-shape func-
tional form. The obtained transition energies are indicated by ar-
rows.

PRB 60
15 769
ELECTRONIC STRUCTURE OF ReS2 AND . . .
TABLE II. Energy positions of various features observed in the EER spectrum of ReS2 and the assign-
ments of interband transitions.
Feature
Energy ͑eV͒
The assignments of interband transitions
A (Ee1x)
B (Ee2x)
C (Eex)
Nonbonding Re 5d t2g→Nonbonding Re 5d t2g
*
*
1.486Ϯ0.005
1.534Ϯ0.005
1.621Ϯ0.005
Nonbonding Re 5d t2g→Nonbonding Re 5d t2g
Nonbonding Re 5d t2g→Sulfur 3p states
D
E
F
G
H
I
J
K
L
1.988Ϯ0.01
2.249Ϯ0.01
2.474Ϯ0.01
2.684Ϯ0.01
2.952Ϯ0.01
3.241Ϯ0.01
3.716Ϯ0.01
4.166Ϯ0.01
4.486Ϯ0.01
4.830Ϯ0.01
5.344Ϯ0.015
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
Re 5d/S 3p bonding→Re 5d/S 3p antibonding
M
N
of ReS2. The calculated energy gap of 1.16 eV agreed well
with our present DOS calculation for the same material. The
electronic structure of ReS2 is found to be determined by
short-range interactions in the S 3p– Re 5d band complex
with the ligand-field splitting of the Re 5d states in the envi-
ronment of the S atoms determining the structure of the d
band. The total and partial density of states were calculated.
Furthermore, ab initio band-structure calculations of Re X2
were also performed by Fang et al.19 The total and partial
DOS of undistorted 3R-ReX2 and distorted Re X2 were, re-
spectively, calculated by the localized-spherical-wave
method.19 From the calculations, the energy gap and the
width of the valence band for the distorted ReS2 ͑ReSe2͒
were determined to be 1.0 eV ͑0.5 eV͒ and 7.05 eV ͑6.8 eV͒,
respectively. The results also showed that the conduction
band is composed mainly of Re 5d states mixed with a small
amount of chalcogen p orbitals. For Re X2, neither the top of
the valence band nor the bottom of the conduction band is at
the ⌫ point. Our results of the total and partial DOS calcu-
lations for ReS2 and ReSe2 using the LAPW are in reason-
able agreement with these reports.
ReS2 and ReSe2 are found to be indirect semiconductors
with an optical gap of 1.37 and 1.19 eV, respectively.16 The
formation of a semiconducting gap is due to the formation of
triclinic low symmetry in the ReS2 and ReSe2 isoelectronic
compounds. These triclinic structures can be considered as
arising from a deformation of the octahedral stacking for-
mula of layered dichalcogenides. The triclinic structure is
derived from the octahedral layered structure by a shift of Re
atoms to form Re4 units in the central plane of the X-M-X
sandwich. Because of the Re-Re interactions, the energy can
be lowered by forming a gap in the middle of bonding and
antibonding t2g bands. The bonding t2g band overlaps
strongly with the X p band and constitutes the valence band;
the antibonding t2g band together with the X p– Re eg hybrid
states form the conduction band. The highest occupied states
around the Fermi level are dxy and dx Ϫy
2
2 orbitals.8,18 The
contribution of chalcogen pz orbitals is much stronger at the
TABLE III. Energy positions of various features observed in the EER spectrum of ReSe2 and the assign-
ments of interband transitions.
Feature
Energy ͑eV͒
The assignments of interband transitions
B (Ee2x)
C (Eex)
Nonbonding Re 5d t2g→Nonbonding Re 5d t2g
*
1.339Ϯ0.005
1.458Ϯ0.005
Nonbonding Re 5d t2g→Selenium 4p states
D
E
F
G
H
I
J
K
L
1.616Ϯ0.01
1.782Ϯ0.01
2.002Ϯ0.01
2.585Ϯ0.01
2.891Ϯ0.01
3.154Ϯ0.01
3.561Ϯ0.01
4.230Ϯ0.01
4.544Ϯ0.01
4.88Ϯ0.01
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
Re 5d/Se 4p bonding→Re 5d/Se 4p antibonding
M
N
5.365Ϯ0.015

15 770
HO, HUANG, CHEN, DANN, AND TIONG
PRB 60
top portion of the valence band than at the bottom portion of
the conduction band. For the px and py orbitals, the top
portion of the valence band has nearly equal contributions
from all chalcogen atoms.
Displayed in Fig. 2 are the ultraviolet photoelectron spec-
tra of the Re 4f core-level electrons for ReS2 and ReSe2. For
both spectra, two peaks are observed and the peak splittings
͑Re 4f7/2 and 4f5/2͒ for ReS2 and ReSe2 have the same value
of 2.4 eV. The energy positions of the two peaks for ReSe2
showed a redshift of ϳ0.8 eV with respect to ReS2 and are,
respectively, located at 40.83 and 43.23 eV. For the Re
metal, the peak positions of the 4f core level are 40.5 and
42.9 eV.20 In comparison with the Re metal, the energy shifts
3/4 of the 4f core levels are most probably influenced by the
bonding of the chalcogen elements with rhenium in the Re X2
compounds. The linewidths of the 4f core levels for Re X2
are estimated to be 0.12 eV for the lower-energy peak and
0.25 eV for the higher one. From the linewidth broadening,
the energy resolutions of our UPS spectra are higher than the
XPS data reported by Fang et al.19
In order to evaluate the electronic band structure, EER
measurements of ReS2 and ReSe2 were also carried out. Four
samples of each material from different growth batches were
chosen. All four samples exhibited similar spectral re-
sponses. The typical EER spectra of ReS2 and ReSe2 at room
temperature in the energy range of 1.3–6 eV are shown in
Figs. 3͑a͒ and 3͑b͒, respectively. In comparison with the
UPS results, more features are detected in the EER spectra
and the derivativelike structures are apparent in the vicinity
of interband transitions. The structures are indicated by ar-
rows and with the letters A–N for ReS2 and B–N for ReSe2
͑as shown in Fig. 3͒. A multitude of interband transitions are
observed as a consequence of the nature of triclinic low-
symmetric structures of Re X2. The amplitudes of two fea-
tures ͑A and B͒ below 1.55 eV for ReS2 and one feature ͑B͒
below 1.45 eV for ReSe2 are much larger than other features.
These features are related to the previously identified band-
FIG. 4. The energy-band-structure schemes of ͑a͒ ReS2 and ͑b͒
ReSe2.
edge excitonic transitions Ee1x and E2ex
.
For ReSe2, only
21
Ee2x is visible at 300 K, whereas for ReS2 both E1ex and Ee2x
need to be included for a good fit of the features. By fitting
the EER spectra with the Aspnes derivative line-shape
expressions,15 we can determine the position of the interband
transitions to an accuracy better than 5 meV. Listed in Tables
II and III are, respectively, the energy positions of interband
transitions for ReS2 and ReSe2 obtained from the EER spec-
tra together with the assignments of the origin of the inter-
band transitions. Comparing the results with those of the
calculated DOS of ReS2 and ReSe2, it is possible to associate
the features of the EER spectra with appropriate interband
transitions. At present, the results may not indicate the exact
location of the critical point transitions associated with each
feature. Nevertheless, this work will prove to be useful in
aiding such identification when a more detailed theoretical
band-structure calculation is available. We can relate the
most prominent features A at 1.486 eV and B at 1.534 eV in
Fig. 3͑a͒, and the feature B at 1.339 eV in Fig. 3͑b͒ as the
band-edge excitons originated from the nonbonding
such an assignment is suggested by the low-temperature pi-
ezoreflectance (P2R) measurements where a series of sharp
features are observed at the energy location above E2ex
.
22
From the EER spectra in Fig. 3, the structures at higher
energies ͑i.e., features D–N for ReS2 and ReSe2͒ are attrib-
uted to interband transitions from states of largely chalcogen
*
X p character to (Re 5d eg )–(X p) hybrid states. With the
results of DOS calculations, EER measurements, and the
optical-absorption measurements, probable band-structure
schemes consistent with Table II for ReS2 and Table III for
ReSe2 are constructed and are shown in Figs. 4͑a͒ and 4͑b͒,
respectively. The band structures of ReS2 and ReSe2 are
composed of three main bands separated by energy gaps.
The lowest band consists of two S 3s or Se 4s bands; the
next bands are mainly S 3p or Se 4p hybridized with Re 5d.
Three Re 5d orbitals form a nonbonding t2g band with dz2 at
the bottom hybridized with S 3p or Se 4p and dx Ϫy2 and dxy
2
*
Re 5d t2g (dxy ,dx2Ϫy2) to 5d t2g transitions. It is noticed that
at the top of the band. The bottom of the conduction band is
composed mainly of Re 5d states mixed with a small amount
of S 3p or Se 4p orbitals. Energy gaps of 1.37 and 1.19 eV
the feature C in Figs. 3͑a͒ and 3͑b͒ is being assigned as the
Re d to X p interband excitonic transition. The reason for

PRB 60
15 771
ELECTRONIC STRUCTURE OF ReS2 AND . . .
separated the top of the valence band and the bottom of the
conduction band for ReS2 and ReSe2, respectively. The an-
bonding Re 5d orbitals. From the results of EER and previ-
ous optical-absorption measurements with appropriate tran-
sition assignments, probable energy-band-structure schemes
for ReS2 and ReSe2, respectively, are constructed.
*
*
tibonding Re 5d t2g band overlaps with the Re 5d eg band.
IV. SUMMARY
ACKNOWLEDGMENTS
In summary, the experimental UPS and EER measure-
ments and theoretical calculation of DOS of ReS2 and ReSe2
have been carried out. A multitude of structures in the UPS
spectra and calculated DOS are observed due to the triclinic
low-symmetry structures of these materials. The contribu-
tions in the top portions of the valence band and the lowest
portions of the conduction band mainly come from the non-
The authors C.H.H. and Y.S.H. acknowledge the support
of the National Science Council of the Republic of China
under Project No. NSC88-2112-M-011-001 and K.K.T. ac-
knowledges the support of the National Science Council of
the Republic of China under Project No. NSC88-2112-M-
019-004.
*
Present address: Department of Electronic Engineering, Kuang
Wu Institute of Technology and Commerce, Peitou, Taipei 112,
Taiwan, Republic of China.
Chem. Solids 60, 1797 ͑1999͒.
12 P. Blaha, K. Schwarz, and J. Luitz, WIEN97—A Full-Potential
Linearized Augmented Plane Wave Package for Calculating
Crystal Properties, Vienna University of Technology, Vienna,
Austria ͑1997͒.
†Author to whom correspondence should be addressed. Electronic
address: ysh@et.ntust.edu.tw
1 J. C. Wildervanck and F. Jellinek, J. Less-Common Met. 24, 73
͑1971͒.
13 M. Cardona, K. L. Shaklee, and F. H. Pollak, Phys. Rev. 154, 696
͑1967͒.
2 S. P. Kelty, A. F. Ruppert, R. R. Chianelli, J. Ren, and M.-H.
Whangboo, J. Am. Chem. Soc. 116, 7857 ͑1994͒.
3 J. V. Marzik, R. Kershaw, K. Dwight, and A. Wold, J. Solid State
Chem. 51, 170 ͑1984͒.
14 Y. S. Huang and Y. F. Chen, Phys. Rev. B 38, 7997 ͑1988͒.
15 D. E. Aspnes, in Handbook on Semiconductors, edited by M.
Balkanski ͑North-Holland, Amsterdam, 1980͒, Vol. 2, p. 109.
16 C. H. Ho, P. C. Liao, Y. S. Huang, T. R. Yang, and K. K. Tiong,
J. Appl. Phys. 81, 6380 ͑1997͒.
4 H. S. Broadbent, L. H. Slangh, and N. L. Jarvis, J. Am. Chem.
Soc. 76, 1519 ͑1954͒.
17 Y. S. Huang, H. M. Chen, C. J. Chang, and G. J. Jan, Chin. J.
Phys. 23, 144 ͑1985͒.
5 S. Harris and R. R. Chianelli, J. Catal. 86, 400 ͑1984͒.
6 B. L. Wheeler, J. K. Leand, and A. J. Bard, J. Electrochem. Soc.
133, 358 ͑1986͒.
18 P. Raybaud, J. Hafner, G. Kresse, and H. Troulhoat, J. Phys.:
Condens. Matter 9, 11 107 ͑1997͒.
7 F. P. Koffyberg, K. Dwight, and A. Wold, Solid State Commun.
30, 433 ͑1979͒.
19 C. M. Fang, G. A. Wiegers, C. Haas, and R. A. de Groot, J. Phys.:
Condens. Matter 9, 4411 ͑1997͒.
8 M. Kertesz and R. Hoffmann, J. Am. Chem. Soc. 106, 3453
͑1984͒.
20 C. D. Wagner, W. M. Riggs, L. E. Davis, J. F. Moulder, and G. E.
Muilenberg, Handbook of X-ray Photoelectron Spectroscopy
͑Perkin-Elmer, Eden Prairie, MN, 1978͒.
9 N. W. Alcock and A. Kjekshus, Acta Chem. Scand. 19, 79
͑1965͒.
21 C. H. Ho, P. C. Liao, Y. S. Huang, and K. K. Tiong, Phys. Rev.
B 55, 15 608 ͑1997͒.
10 H.-J. Lamfers, A. Meetsma, G. A. Wiegers, and J. L. de Boer, J.
Alloys Compd. 241, 34 ͑1996͒.
22 C. H. Ho, Y. S. Huang, K. K. Tiong, and P. C. Liao, J. Phys.:
Condens. Matter 11, 5367 ͑1999͒.
11 C. H. Ho, Y. S. Huang, P. C. Liao, and K. K. Tiong, J. Phys.