Vibrational Spectroscopy 37 (2005) 33–38

Size effects in the Raman spectra of TiO2 nanoparticles

Hyun Chul Choi, Young Mee Jung, Seung Bin Kim*
Department of Chemistry, Pohang University of Science and Technology, San 31, Hyojadong, Namgu, Pohang 790-784, Republic of Korea
Received 10 October 2003; received in revised form 24 May 2004; accepted 26 May 2004
Available online 8 September 2004

Abstract

To investigate the relationship between the particle size and the Raman bands of TiO2 nanoparticles, two different size-selected samples of
TiO2 nanoparticles were investigated using transmission electron microscopy (TEM), X-ray absorption near edge structure (XANES),
extended X-ray absorption fine structure (EXAFS), and Raman spectroscopy. Analysis of the XANES spectra showed that the nanoparticles in
both sample A (particle size
30 nm) and sample B (particle size
12 nm) have an anatase structure, and that there is no difference between
the electronic structure of titanium atoms in TiO2 nanoparticles and that of titanium atoms in bulk TiO2. The EXAFS results indicate that a
contraction effect occurs in the anatase TiO2 nanoparticles. In the Raman spectra, both broadening and shifts of the Raman bands with
decreasing particle diameter were observed. In this paper, these Raman shifts are attributed to the effects of decreasing particle size on the
force constants and vibrational amplitudes of the nearest neighbor bonds.
# 2004 Elsevier B.V. All rights reserved.

Keywords: Raman spectra; TiO2; Nanoparticles; X-ray absorption spectroscopy (XAS)

1. Introduction the lowest-energy spherical mode of a free particle, corre-

Nanometer-sized materials have recently gained a con-
siderable amount of attention because of their unique phy-
sical and chemical properties and their importance in
technological applications [1,2]. Because of their large sur-
face-to-volume ratios and quantum-size effect, their proper-
ties and structural stability are very different to those of their
bulk counterparts [3–5]. A significant motivation of current
research into nanometer-sized materials is the need to
develop an understanding of the relationships between their
structures and their properties. Despite the extensive research
efforts in this area using various characterization techniques,
such an understanding has still not been achieved.
Raman spectroscopy is a powerful method for the inves-
tigation of the structural properties of nanoparticles because
the variations in Raman spectra with decreases in particle size
can be easily detected. Although there is still some contro-
versy as to what effects variation in particle size has on
phonon modes and on the electron-phonon interaction [6,7],
Raman studies of these materials are useful in two distinct
ways. Firstly, it is possible to determine the size of nano-
particles from a measurement of the maximum of the low-
frequency Raman band [8–10]. The frequency n (in cm1) of
*
Corresponding author. Tel.: þ82-54-279-2106; fax: þ82-54-279-3399.
E-mail address: sbkim@postech.edu (S.B. Kim).
sponding to angular momentum l ¼ 0, is given by [11,12]
0:7vL
n¼ (1)
dc
where vL is the speed of the longitudinal sound waves, c is
the vacuum light velocity of sound, and d is particle diameter.
Secondly, this technique provides valuable information about
the surface modes and thus about the effects of the finite size
of the nanoparticles, because the surface modes become
dominant as particle size decreases due to the increase in
the surface-to-volume ratio, which means that new bands can
be observed. For example, Campbell and Fauchet have
reported that grain size effects produce large shifts and
broadening of the Raman spectrum of silicon [13]. Palkar
et al. studied the Raman spectra of Si-doped PbTiO3 and
reported that the broadening and shifts of the Raman bands of
PbTiO3 with increasing Si content can be attributed to an Si-
induced finite size effect [14]. Zuo et al. obtained the Raman
spectra of SnO2 nanoparticles and reported that two new
characteristic bands resulting from surface phonon modes are
present when the particle diameter is less than 10 nm [15].
The dependence of Raman bands on particle size has also
been observed for MgO, ZnO, CdS, and other nanoparticles
[16–19].
In contrast to the above-mentioned results for SnO2
nanoparticles, only band broadening and shifts of the Raman

0924-2031/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.vibspec.2004.05.006
34 H.C. Choi et al. / Vibrational Spectroscopy 37 (2005) 33–38
bands in the spectra of TiO2 nanoparticles with decreasing
particle size have been observed, even though TiO2 and
SnO2 possess the same crystal structure (tetragonal) [20–23].
Parker et al. have suggested that the observed variations in
the Raman spectra with particle size are related to the
oxygen stoichiometry of the TiO2 nanoparticles and are
not due to any internal stress or grain size effects [20].
However, the Raman results described above show that grain
size effects do manifest within nanoparticles. Further, Luca
et al. detected the presence of contracted Ti–O bonds in
anatase TiO2 nanoparticles using X-ray absorption spectro-
scopy (XAS) [24]. Toneje et al. reported that a dependence
of the lattice parameter on grain size arises for anatase TiO2
nanoparticles [25]. Clearly an improved understanding of
the origin of the observed variation of Raman bands of TiO2
nanoparticles with particle size is required.
The positions and widths of Raman bands of materials are
related to their vibrational and structural properties. Thus, a
better understanding of the Raman spectra of a material can
be obtained by examining the local structure and electronic
environment of its constituent atoms. XAS is known to
provide valuable information about the local structure around
atoms of interest [26,27]. The absorption peak features in the
X-ray absorption near edge structure (XANES) can be used
to extract useful structural parameters such as the oxidation
states of the chemical species, site symmetry, and covalent
bond strengths. Further, the study of extended X-ray absorp-
tion fine structure (EXAFS) is a powerful technique for the
local structural analysis of short-range ordered compounds
because it is very sensitive to the local atomic environment
and does not depend on long-range order. Therefore, a
combined analysis of Raman and XAS spectra can greatly
enhance the understanding of the experimental results.
The objective of the present study is to elucidate the origin
of the variations in the Raman bands of TiO2 nanoparticles
with particle size. To achieve this objective, two samples of
size-selected TiO2 nanoparticles were prepared and each of
these samples was investigated using X-ray absorption
spectroscopy and Raman spectroscopy. Both broadening
and shifts of the Raman bands with decreasing particle
diameter were observed. In this paper, we conclude that
these Raman shifts are due to the effects of decreasing
particle size on the force constants and vibrational ampli-
tudes of the nearest neighbor bonds.
2. Experimental
TiO2 nanoparticles were prepared via a hydrolysis method
using titanium isopropoxide [Ti[OCH(CH3)2]4, Aldrich,
99.999%] as the starting material without further purifica-
tion. The Ti[OCH(CH3)2]4 solution was added to 200 mL of
2-propanol at room temperature with vigorous stirring, and
then 800 mL of deionized water was added to the solution.
The reaction conditions were maintained for 1 h, which was
assumed to be sufficient time for the reactions, hydrolysis,
and particle formation to reach equilibrium. Sample A was
obtained by heating the precipitated particles at 450 8C for
1 h. Sample B was obtained by adjusting the stirring rate of
the original reaction mixture (i.e. Ti[OCH(CH3)2]4 and 2-
propanol) to 200 rpm, and the precipitated particles were
dried with a freeze-drier. The TiO2 nanoparticles of sample
B were then obtained by heating the dried particles at 450 8C
for 1 h.
TEM images were obtained using a Phillips CM20T/
STEM Electron Microscope at an accelerating voltage of
200 kV. FT-Raman spectra were obtained at room tempera-
ture using a Bruker FRA 106/S FT-Raman spectrometer
(Germany) equipped with an InGaAs detector. The radiation
from an Nd-YAG laser (1.5 W) was used as the excitation
source. Ti K-edge X-ray absorption spectra were recorded
using the 7C beam line at the Pohang Light Source (PLS)
with a ring current of 120
170 mA at 2.5 GeV. A Si(1 1 1)
monochromator crystal was used with detuning to 85% in
intensity to eliminate high-order harmonics. Data were
collected in transmission mode using gas-filled ionization
chambers as detectors. Energy calibration was carried out
for all measurements using Ti foil placed in front of the third
ion chamber, assigning the first inflection point to 4966 eV.
The measured X-ray absorption spectra below the pre-
edge region were fitted to a straight line, and then the
background contribution above the post-edge region,
m0 ðEÞ, was fitted to a high order polynomial (cubic spline).
The fitted polynomials were extrapolated over the entire
energy region and subtracted from the total absorption
spectra. The absorption spectra with the background sub-
tracted were normalized for the above energy region,
wðEÞ ¼ fmðEÞ  m0 ðEÞg=m0 ðEÞ: The normalized k3-weight-
ed EXAFS spectra, k3 wðkÞ, were Fourier transformed to
reveal the contribution of each bond pair to the Fourier
transform (FT) peak.
3. Results and discussion
Fig. 1 shows the TEM images of the samples. It is obvious
from these images that the samples are disperse, with an
average particle size of
30 nm for sample A, and
12 nm
for sample B. The particle sizes of samples A and B
determined from the TEM images are in good agreement
with the values obtained by applying the Scherrer equation
to the XRD patterns (results not shown) [28]. The particle
sizes of samples A and B determined from the TEM images
are in good agreement with the values determined from the
XRD analysis [29]. The TEM results show that size-selected
samples of TiO2 nanoparticles have been successfully pre-
pared from the Ti[OCH(CH3)2]4 precursor. To elucidate the
variations in structure and electronic state of the titanium
ions with decreasing particle size, XAS data were collected
for the two samples of TiO2 nanoparticles.
Absorption in the XANES region varies with the site
symmetry, coordination, and the oxidation states of the
 H.C. Choi et al. / Vibrational Spectroscopy 37 (2005) 33–38 35

central atoms, so the XANES has been widely used to

characterize various forms of TiO2 [24,30–33]. A Ti K-edge
XANES spectrum is usually divided into two regions. The
region above approximately 4980 eV is assigned to the well-
understood dipole-allowed 1s ! np (n 4) transitions;
below 4980 eV lies the so-called pre-edge peaks. In previous
studies of TiO2 [24,31,32], four pre-edge peaks (usually
labeled A1, A2, A3, and B) were observed in the range 4960–
4980 eV. Feature A2 is sometimes observed as a weak
shoulder on the low energy side of peak A3, and requires
good energy resolution to be observed. The pre-edge peak
arises from transitions of electrons to bound excited elec-
tronic states, but this electronic excitation is strongly modu-
lated by the surrounding atoms in the short to medium range
(<100 absorbing atoms) [24]. Therefore the pre-edge region
contains much potentially useful structural and electronic
information, and is frequently used to elucidate the structure
of TiO2 powders in the Ti K-edge XANES spectrum.
Fig. 2 shows the Ti K-edge XANES spectra of the two
samples of TiO2 nanoparticles and of a reference material
(anatase TiO2, Aldrich, 99.9%). Based on previous band
structure calculations [34,35], peaks A3 and B in the Ti K-
edge XANES spectra were assigned to 1s transitions to the
t2g and eg bands, which do not have pure Ti d character
because of the hybridization of Ti 3d and 4p orbitals on
different sites in the conduction band region [24,33]. How-
ever, the assignment in the literature of the Al and A2
transitions has not been unanimous. Parlebas et al. suggested
that peak A1 is essentially due to the quadruple transition to
3d states of t2g symmetry on the central atom, which are
Fig. 1. TEM images of (a) sample A and (b) sample B. shifted down from their unperturbed position by the 1s core-
hole potential (a type of excitonic mechanism) [32]. In
contrast, Wu et al. based their assignment on multiple
scattering calculations, and concluded that peaks A1, A2,
Normalized Intensity (arb. unit)

A1 A2 A3 B

A1A2 A3 B

(c)

(b)
4968 4970 4972 4974 4976
(a)
Photon energy (eV)

4960 4970 4980 4990 5000 5010 5020

Photon energy (eV)

Fig. 2. Ti K-edge XANES spectra of (a) reference (anatase TiO2), (b) sample A, and (c) sample B.
36
and A3 arise dominantly from the hybridization of 3d-4p
orbitals, and reflect a different degree of hybridization with
the central Ti 4p orbitals due to the different distance of the
two Ti shells [33].
Because XANES measurements involve the excitation of
electrons from core levels to partially filled and empty states,
the peak positions and spectral features of the Ti K-edge
XANES spectra will be affected not only by the oxidation
state but also by other chemical features of Ti species, such
as the structural symmetry and the covalent/ionic character
of the bonds between Ti and neighboring atoms. In Fig. 2 it
can be seen that the main features of the spectra of the TiO2
nanoparticles show no noticeable variation with decreases in
particle size. This indicates that the nanoparticles in samples
A and B have anatase structures, and that there is no change
in the oxidation state of the titanium ions with decreasing
particle size. However, when the particle size is decreased
from 30 to 12 nm, the ratio of the intensity of A2 to that of A3
increases slightly. This increase can be attributed to the
decrease in particle size. It is known that peak A2 arises
predominantly from the surface structure of anatase particles
[24]. As the particle size decreases, peak A2 becomes
dominant due to the increase in the surface-to-volume ratio.
These observations are also in good agreement with a recent
study of the anatase Ti K-edge XANES [24]. Thus, our
XANES results show that the nanoparticles in samples A and
B have anatase structures, and that there is no difference
between the oxidation state of the titanium ions in the
nanoparticles and that of titanium ions in bulk TiO2.
When the particle size decreases to the nanometer scale,
the structural properties of nanoparticles can become dif-
ferent to those of the bulk material. Bulk anatase has a
tetragonal structure (space group I41/amd) containing 12
atoms per unit cell with lattice parameters a ¼ 3.784 Å and
20
Fourier transform magnitude (a. u.)
16
12
0
0 2
Fig. 3. Fourier transforms (FTs) of the k3-weighted EXAFS spectra.
H.C. Choi et al. / Vibrational Spectroscopy 37 (2005) 33–38
c ¼ 9.514 Å [36]. The Ti4þ ions are surrounded by six O2
(4  Ti–O ¼ 1.939 Å, 2  Ti–O ¼ 1.980 Å) at the corners
of a slightly distorted octahedron, and each O2 is sur-
rounded by three Ti4þ lying in a plane at the corners of an
equilateral triangle. The anatase structure comprises two
interpenetrating zigzag chains of edge-shared octahedra,
which are linked to form a three-dimensional network of
edge-shared octahedra. To elucidate the variation in the local
structure around Ti atoms as a function of particle size, we
carried out an EXAFS analysis.
Fig. 3 shows the EXAFS spectra of the anatase TiO2
nanoparticles and that of the reference material. The Fourier
transforms alter the phase shift parameters, so the peak
positions are shifted to values below those in the original
spectrum. In the spectrum of the reference TiO2, the FT peak
at
1.6 Å is due to the single scattering path of Ti ! O by
six-coordinated oxygen nearest neighbors around the Ti
atom. On the other hand, the two peaks in the range 2–
4 Å are due to the single scattering path of Ti ! Ti by
neighboring Ti atoms. The magnitudes of the FT at
2.7 and

3.6 Å correspond to single scattering by the edge-shared
octahehedra and corner-shared octahedra respectively. The
magnitudes of the FT above
4.0 Å indicate the single and
multiple scattering contributions. The general features of the
spectra of the anatase TiO2 nanoparticles have a strong
resemblance to those of the spectrum of the bulk, with a
reduction in the magnitude of the FT peaks with decreasing
particle size. It is also observed that the FT peaks in the
range 2–4 Å are slightly shifted down compared to those of
the reference material. The first observation can be
explained in terms of the effects of decreasing particle size:
because the surface-to-volume ratio and the number of
distorted sites at the surface increase as the particle size
decreases, a reduction in the magnitude of the FT peaks of
Sample B (12 nm)
Sample A (30 nm)
Reference TiO2
4 6 8 10
Interatomic distance (Å)
 H.C. Choi et al. / Vibrational Spectroscopy 37 (2005) 33–38 37

Eg

Raman intensity (a. u.)

Eg
A1g B1g Eg
(c)

(b)

(a)

800 700 600 500 400 300 200 100

-1
Wavenumber (cm )
Fig. 4. Raman spectra of (a) reference, (b) sample A, and (c) sample B.

nanoparticles is generally observed [37]. The second obser-
vation suggests that the size-induced radial pressure of TiO2
nanoparticles leads to a volume contraction as the particle
size decreases. The next nearest neighbor (Ti–Ti) shells thus
experience an increase in the mean square relative displace-
ment (MSRD) with decreases in particle size, because the
magnitude of the FT peak of a bond pair is inversely related
to the Debye–Waller factor corresponding to the MSRD of
the interatomic distance; similar observations have been
reported for TiO2 and CdS nanoparticles [24,38].
TiO2 nanoparticles have frequently been investigated with
Raman spectroscopy because of the unusual band broad-
ening and shifts of Raman bands with decreasing particle
size. However, there is no general agreement about the
origin of the broadening and shifts of the Raman bands.
In contrast, Xu et al. have tried to explain the variation in the
Raman bands with a phonon confinement model [23]. On the
basis of the Heisenberg uncertainty principle, the relation-
ship between particle size and phonon position can be
expressed as follows:
2
h the anatase phases of the nanoparticles of samples A and B
DX DP
4 possess a certain degree of long-range order. Comparing the
where DX is the particle size, DP is the phonon momentum
distribution, and 
h is the reduced Planck’s constant. As the
particle size decreases, the phonon is increasingly confined
within the particle and the phonon momentum distribution
increases. This broadening of the phonon momentum leads
to a broadening of the scattered phonon momentum accord-
ing to the law of conservation of momentum. This phonon
dispersion causes asymmetric broadening and may lead to a
shift of the Raman bands [23].
According to factor group analysis, anatase has six Raman
active modes (A1g þ 2B1g þ 3Eg). The Raman spectrum of
an anatase single crystal has been investigated by Ohsaka,
who concluded that the six allowed modes appear at
144 cm1 (Eg), 197 cm1 (Eg), 399 cm1 (B1g), 513 cm1
(A1g), 519 cm1 (B1g), and 639 cm1 (Eg) [39]. In this study,
we assigned and interpreted the Raman bands of the TiO2
nanoparticles using earlier results obtained for the bulk
phase. Fig. 4 shows the Raman spectra of the two samples
of TiO2 nanoparticles and of the reference spectrum of
anatase TiO2. In order to see the differences between the
spectra more clearly, the wavenumbers and the full-widths at
half-maximum (FWHM) of the bands are given in Table 1.
The main features of the spectra of samples A and B are
very similar to those of the reference TiO2, which means that
(2)
three Raman spectra, it is clear that the Raman bands shift
towards higher wavenumber and their intensities relatively
decrease as the particle size decreases. However, the band

Table 1
The wavenumber and the FWHM (in parentheses) of the samples (cm1)

Eg Eg B1g A1g Eg

Reference TiO2 143 (8.61) 196 (4.28) 396 (18.34) 514 (21.93) 637 (23.36)
Sample A 144 (13.23) 196 (4.39) 396 (36.53) 515 (23.87) 638 (25.01)
Sample B 145 (11.31) 196 (9.14) 398 (24.20) 516 (30.68) 639 (31.73)
38 H.C. Choi et al. / Vibrational Spectroscopy 37 (2005) 33–38
broadening is insignificant, and so is not considered further References
in this study. These observations cannot be explained in
terms of changes in oxygen stoichiometry with decreasing [1]
particle size, because the XANES results show that there is [2]
[3]
no variation of the electronic structure around the Ti atoms.
[4]
Thus, the observed shift is due to the effect of decreasing
particle size on other properties of the nanoparticles. When [5]
the particle size decreases to the nanometer scale, two [6]
effects on the vibrational properties of these materials might [7]
occur. Firstly, a volume contraction occurs within the nano-
[8]
particles that is due to the size-induced radial pressure,
which leads to increases in the force constants as a result [9]
of the decreases in the interatomic distances. In vibrational
transitions, the wavenumber varies approximately in propor- [10]
tion to k1/2, where k is the force constant. Consequently, the
[11]
Raman bands shift towards a higher wavenumber due to the [12]
increasing force constants. Secondly, the contraction effect [13]
induces decreases in the vibrational amplitudes of the near- [14]
est neighbor bonds due to the increase of the MSRD, which
[15]
can be interpreted as a measure of the static disorder and
thermal vibrational disorder of a material [38,40]. This [16]
decrease in vibrational amplitude with decreasing particle [17]
size affects the intensity of the Raman bands. We conclude
that the variation in the Raman spectra of TiO2 nanoparticles [18]
is due to the effect of smaller particle size on the force
[19]
constants and vibrational amplitudes of the nearest neighbor [20]
bonds. [21]
[22]
4. Conclusion [23]
[24]
[25]
The broadening and shifts of Raman bands of anatase [26]
TiO2 nanoparticles with decreasing particle size were inves-
tigated using X-ray absorption spectroscopy and Raman [27]
spectroscopy. The XANES results show that the prepared
[28]
30 and 12 nm TiO2 nanoparticles have an anatase structure, [29]
and that there is no difference between the oxidation state of
the titanium ions in the nanoparticles and that of titanium [30]
ions in bulk TiO2. The EXAFS results indicate that size-
induced radial pressure in TiO2 nanoparticles increases as [31]
the particle size decreases, and this results in a volume
contraction. Additionally, the next nearest neighbor (Ti–Ti) [32]
shells are strongly influenced by decreases in the particle
size, which produces increases in the mean square relative [33]
displacement (MSRD). Both broadening and shifts of the
[34]
Raman bands were observed with decreasing particle dia- [35]
meter. The observed Raman shifts are thus due to the effects
of decreasing particle size on the force constants and [36]
vibrational amplitudes of the nearest neighbor bonds.
[37]
Acknowledgements [38]
The present study was supported by the Brain Korea 21 [39]
project. Y.M. Jung acknowledges the support of the Korea [40]
Research Foundation (KRF-2002-015-CP0164).
A. Henglein, Chem. Rev. 89 (1989) 1861.
H. Weller, Angew. Chem. Int. Ed. Engl. 32 (1993) 41.
A.P. Alivisatos, J. Phys. Chem. 100 (1996) 13226.
C.-C. Chen, A.B. Herhold, C.S. Johnson, A.P. Alivisatos, Science
276 (1997) 398.
S.K. Sinha, Appl. Surf. Sci. 182 (2001) 176.
G. Scamarcio, M. Lugará, D. Manno, Phys. Rev. B 45 (1992) 13792.
A.M. de Paula, L.C. Barbosa, C.H.B. Cruz, O.L. Alves, J.A.
Sanjurio, C.L. Cesar, Appl. Phys. Lett. 69 (1996) 357.
A. Turković, M. Ivanda, S. Popović, A. Tonejc, M. Gotić, P. Dubček,
andS. Musić, J. Mol. Struct. 410–411 (1997) 271.
M. Gotić, M. Ivanda, S. Popović, S. Musić, A. Sekulić, A. Turković,
K. Furić, J. Raman Spectrosc. 28 (1997) 555.
S. Musić, M. Gotić, M. Ivanda, S. Popović, A. Turković, R. Trojko,
A. Sekulić, K. Furić, Mater. Sci. Eng. B 47 (1997) 33.
H. Lamb, Proc. Lond. Math. Soc. 13 (1982) 187.
A. Tamura, K. Higeta, T. Ichinokawa, J. Phys. C 15 (1982) 4895.
I.H. Campbell, P.M. Fauchet, Solid State Commun. 58 (1986) 739.
V.R. Palkar, S.C. Purandare, P. Ayyub, R. Pinto, J. Appl. Phys. 87
(2000) 462.
J. Zuo, C. Xu, X. Liu, C. Wang, C. Wang, Y. Hu, Y. Qian, J. Appl.
Phys. 75 (1994) 1835.
R.G. Schlecht, H.K. Böckelmann, Phys. Rev. Lett. 31 (1973) 930.
J. Zuo, C. Xu, L. Zhang, B. Xu, R. Wu, J. Raman Spectrosc. 32
(2001) 979.
B. Schreder, C. Dem, M. Schmitt, A. Materny, W. Kiefer, U.
Winkler, E. Umbach, J. Raman Spectrosc. 34 (2003) 100.
S. Hayashi, H. Kanamori, Phys. Rev. B 26 (1982) 7079.
J.C. Parker, R.W. Siegel, Appl. Phys. Lett. 57 (1990) 943.
W. Ma, Z. Lu, M. Zhang, Appl. Phys. A 66 (1998) 621.
A. Gajović, M. Stubičar, M. Ivanda, K. Furić, J. Mol. Struct. 563–
564 (2001) 315.
C.Y. Xu, P.X. Zhang, L. Yan, J. Raman Spectrosc. 32 (2001) 862.
V. Luca, S. Djajanti, R.F. Howe, J. Phys. Chem. B 102 (1998) 10650.
A.M. Tonejc, I. Drerdj, A. Tonejc, Mater. Sci. Eng. C 19 (2002) 85.
B.K. Teo, EXAFS: Basic Principles and Data Analysis, Springer-
Verlag, New York, 1986.
B.K. Teo, D.C. Joy, EXAFS Spectroscopy, Plenum Press, New York,
1981.
X. Ye, J. Sha, Z. Jiao, L. Zhang, Nanostruct. Mater. 8 (1997) 919.
H.C. Choi, H.-J. Ahn, Y.M. Jung, H.J. Shin, S.B. Kim, Y.-E. Sung,
M.K. Lee, Appl. Spectrsc. 58 (2004) 598.
R. Brydson, H. Sauer, W. Engel, J.M. Thomas, E. Zeitler, N. Kosugi,
H. Kuroda, J. Phys. Condens. Matter 1 (1989) 797.
G. Antonioli, D. Bersani, P.P. Lottici, I. Manzini, G. Gnappi,
A. Montenero, Nucl. Instrum. Methods Phys. Rev. B 97 (1995)
198.
J.C. Parlebas, M.A. Khan, T. Uozumi, K. Okada, A. Kotani, J.
Electron Spectrosc. Relat. Phenom. 71 (1995) 117.
Z.Y. Wu, G. Ouvrard, P. Gressier, C.R. Natoli, Phys. Rev. B 55
(1997) 10382.
L.A. Grunes, Phys. Rev. B 27 (1983) 2111.
M.A. Khan, A. Kotani, J.C. Parlebas, J. Phys. Condens. Matter 3
(1991) 1763.
C.J. Howard, T.M. Sabine, F. Dickson, Acta Crystallogr. B 47 (1991)
462.
H.C. Choi, S.Y. Lee, S.B. Kim, M.G. Kim, M.K. Lee, H.J. Shin, J.S.
Lee, J. Phys. Chem. B 106 (2002) 9252.
J. Rockenberger, L. Troger, A. Kornowski, T. Vossmeyer, A.
Eychmuller, J. Feldhaus, H. Weller, J. Phys. Chem. B 101 (1997)
2691.
T. Ohsaka, J. Phys. Soc. Jpn. 48 (1980) 1661.
M.A. Marcus, M.P. Andrews, J. Zegenhagen, A.S. Bommannavar, P.
Montano, Phys. Rev. B 42 (1983) 3312.