Materials Chemistry and Physics 112 (2008) 146–153
Contents lists available at ScienceDirect
Materials Chemistry and Physics
journal homepage: www.elsevier.com/locate/matchemphys
Structural and photocatalytic properties of iron- and europium-doped TiO
2
nanoparticles obtained under hydrothermal conditions
L. Diamandescu
a,∗
, F. Vasiliu
a
, D. Tarabasanu-Mihaila
a
,M.Feder
a
, A.M. Vlaicu
a
,
C.M. Teodorescu
a
, D. Macovei
a
, I. Enculescu
a
, V. Parvulescu
b
, E. Vasile
c
a
National Institute of Materials Physics, Atomistilor 105 bis, P.O. Box MG-7, Bucharest, Romania
b

2
O. The structure, morphology and optical
peculiarities were investigated by means of X-ray diffraction (XRD), transmission electron microscopy
(TEM), extended X-ray absorption fine structure (EXAFS), M
¨
ossbauer spectroscopy and UV–vis mea-
surements. The photocatalytic performance was analysed in the photodegradation reaction of phenol.
Rietveld refinements of XRD patterns reveal that the as-prepared samples consist in iron- and europium-
doped TiO
2
in the tetragonal anatase structural shape, with particle size as low as 15 nm. By means of
M
¨
ossbauer spectroscopy on both
57
Fe and
151
Eu isotopes as well as by EXAFS analyses, the presence of
Fe
3+
and/or Eu
3+
ions in the nanosized powders has been evidenced. It was found that iron and europium
ions can substitute for titanium in the anatase structure. From the UV–vis reflection spectra, by using
the transformed Kubelka–Munk functions, the band gap energy (E
g
) of the hydrothermal samples has
been determined in comparison with that of Degussa P-25 photocatalyst. A decrease of E
g
from 2.9 eV

codoped with Fe
3+
and Eu
3+
by sol–gel method, as compared with
undoped or monodoped TiO
2
nanoparticles [16].
It is already established that material properties depend
strongly on precursors and synthesis methods in correlation with
the thermodynamic process parameters. For the synthesis of
nanoparticle systems the hydrothermal method was intensively
utilised in the last decade [17–21]. However, no reports on the
hydrothermal synthesisof iron- and europium-codoped TiO
2
mate-
rials have been published, by our knowledge.
It is the aim of this work to present the hydrothermal synthesis
of iron- and europium-doped and -codoped TiO
2
nanoparticle sys-
tems, their microstructure, morphology and catalytic properties in
the photodegradation of phenol, in both UV and visible light region.
0254-0584/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.matchemphys.2008.05.023
L. Diamandescu et al. / Materials Chemistry and Physics 112 (2008) 146–153 147
2. Experimental
2.1. Hydrothermal synthesis
Fe
3+

C was about 15 atm.
The autoclave was cooled to room temperature and the resulting colloidal suspen-
sion of iron/europium-doped titanium oxide was heated at 50
◦
C for several hours
to remove water. The following hydrothermal sample codes are used in this paper:
undoped sample (TiO
2
); TiO
2
: 1 at.% Fe (sample TF); TiO
2
: 0.5 at.% Eu (sample TE);
TiO
2
: 1 at.% Fe, 0.5 at.% Eu(sample TFE). Owing to the nature of the synthesis process,
the accurate doping level could not be consistently predicted and the percent values
represent the nominal values of Fe or Eu atomic concentrations.
2.2. Structural characterisation
X-ray diffraction (XRD) patterns obtained on DRON 2 X-ray diffractome-
ter (linked to a data acquisition and processing facility with CuK
␣
radiation
 =1.540598
˚
A and a graphite monochromator) were used to determine the iden-
tity of any phase present and their crystallite size. JEOL 200 CX electron microscope
operating at an accelerating voltage of 200 kV was utilised to obtain information
about the structure and morphology of mixed oxide nanoparticles. Particle sizes
were measured from bright field (BF) and dark field (DF) images, whereas the

2
O
3
standards. The TiO
2
(anatase) standard has been measured
correspondingly at theTi K-edge. Normalized EXAFS function (k)(k = photoelectron
wave number) was calculated after subtraction of pre-edge and post-edge smooth
backgrounds, (fitted by Victoreen formula and cubic splines, respectively) from the
absorption spectra. k
n
-weighted (k)(n = 2, 3) was Fourier transformed over the k-
range 2.8–11.2
˚
A
−1
(Fe K) or 1.3–10.2
˚
A
−1
(Eu L
2
), to acquire preliminary information
on the dopant environment. Radial ranges of interest in the Fourier transforms (FT)
were further isolated by Hanning-function windows, backtransformed into k-space,
and non-linearly fitted by a least-square method. The fit provided the interatomic
distances and coordination numbers in the close neighbouring shells of the absorb-
ing atoms (Fe, Eu). The photoelectron backscattering amplitudes and phases, as
calculated by the FEFF6 code [29], have been used in the fitting runs. The UV–vis
measurements have been performed on the PerkinElmer Lambda 45 Spectrome-

doping ions in the TiO
2
lattice. A particle mean size of about 15 nm
(calculated with Scherrer equation) was found to characterise the
hydrothermally doped anatase.
3.2. Transmission electron microscopy
The electron diffraction and electron microscopy analysis evi-
dence the presence of anatase like structure with particle mean size
as low as 15 nm and a strong morphology dependence on doping
element. Thus, a rectangular and quadratic morphology is predom-
Fig. 1. XRD patterns of hydrothermally synthesized doped and undoped TiO
2
.
148 L. Diamandescu et al. / Materials Chemistry and Physics 112 (2008) 146–153
Fig. 2. TEM images of undoped TiO
2
(a), Eu-doped TiO
2
(b) and Fe- and Eu-doped TiO
2
(c).
inant for undoped and Fe-doped TiO
2
samples (Fig. 2a). In Eu-doped
TiO
2
many elongated particles can be seen (Fig. 2b) whereas large
shape diversity is found in codoped Fe and Eu-TiO
2
specimen

ple, 90% enriched in
57
Fe in order to improve the signal-to-noise
ratio (Fig. 5). The best fit was obtained with the isotropic elec-
tronic relaxation model [23]. The values of M
¨
ossbauer hyperfine
parameters (isomer shift and quadrupole splitting) are character-
istics for Fe
3+
[22]. In others words, the iron ions are present in the
TiO
2
lattice, the spin–spin electron interaction between the neigh-
bouring Fe
3+
ions giving rise to the relaxation M
¨
ossbauer spectrum
[23].
3.4. EXAFS
A more detailed analysis on the dopant location and resulting
interactions has been carried out by EXAFS investigations. The k
3
-
weighted Fe K-edge EXAFS spectra of the doped samples TF, TFE and
␣-Fe
2
O
3

transforms rules out the iron segregation to an oxidized Fe
2
O
3
phase. The photocatalyst FTs closely resemble that of TiO
2
, sug-
gesting similar Fe and Ti surroundings in the doped and undoped
samples. In order to verify this statement, the filtered EXAFS of
L. Diamandescu et al. / Materials Chemistry and Physics 112 (2008) 146–153 149
Fig. 3. (a) HRTEMimage showing (10 1)anatase planesin a TiO
2
nanocrystal belong-
ing to (Fe, Eu)-doped specimen; (b) SAED pattern of the same specimen showing the
main diffraction rings of polycrystalline anatase.
Fig. 4.
151
Eu M
¨
ossbauer spectrum of the sample TiO
2
: 1 at.% Fe, 0.5 at.% Eu showing
the presence of Eu
3+
ions in the host lattice.
Fig. 5.
57
Fe M
¨
ossbauer spectrum of the sample TiO

˚
A,
up to values close to the Fe
O bond length in ␣-Fe
2
O
3
, while the
metal–metal distances remain unchanged (sample TF) or shorten
(TFE) with respect to these distances in anatase structure [24].A
similar effect was recently reported by Zhu et al. [25] in Fe-doped
anatase samples, being explained by changes of the Ti
O Ti bond
angles after the Ti substitution by Fe.
Europium environment in the sample TFE was investigated by
EXAFS at the Eu L
2
-edge (7617 eV). Since Eu L
3
-edge EXAFS has
stronger oscillations it could not be analysed, due to the superposi-
tion of the Fe K-edge (7112 eV) at 135 eV above Eu L
3
(6977 eV). The
k
2
-weighted EXAFS of the sample is shown in Fig. 7a, together with
the spectra of Eu
2
O

2
O
3
, its maximum at 2.83
˚
A seems closer related
Table 1
Fe and Eu local environments (interatomic distances R, coordination numbers N)in
the doped samples (TF, TFE), as inferred by the fit of EXAFS
Sample Reference atom R (
˚
A)/N
␣-Fe
2
O
3
Fe 2.03/6 O 2.95/4 Fe
3.38/3 Fe, 3 O
3.69/6 Fe, 6 O
Eu
2
O
3
(cubic) Eu 2.34/6 O 3.60/6 Eu
TiO
2
(anatase) Ti 1.95/6 O 3.04/4 Ti
TiO
2
(rutile) Ti 1.96/6 O 2.96/2 Ti

˚
A resulted in ∼6Oand2Ti
atoms around Eu, at distances of 2.33
˚
A and 3.37
˚
A, respectively (see
Table 1). The presence of titanium in the europium neighbourhood
emphasizes the Eu accommodation on Ti sites in the TiO
2
lattice,
similarly with the Fe case. The large Eu
3+
ions (0.947
˚
A) locally
expand the host structure, with elongations of the interatomic dis-
tances almost equalling with the difference between Eu
3+
and Ti
4+
ionic radii (0.34
˚
A).
A peculiar effect of the Eu incorporation into the TiO
2
lattice
is the change of the local symmetry around Eu, from anatase to
rutile structure. This is indicated by the lowering of the number of
the next-nearest Ti neighbours from four, specific to anatase struc-

/(2R), [27]. Using the Tauc plot (F(R) h)
n
vs
h) [28], where h is the photon energy and n = ½ for direct band
gap semiconductors, the band gap energies were deduced from
the intersection of the Tauc’s region extrapolation with the pho-
ton energy axis (Fig. 9). The calculated band gap energy values are
given in Table 2. For the hydrothermal TiO
2
undoped sample and
TE sample, the band gap energy is about 3.0(3) eV, which is higher
than 2.9(7) eV obtained for the TiO
2
Degussa P-25 sample. The low-
est band gap energy of 2.84 eV was observed for TF sample, while
for TFE sample the energy was 2.92 eV.
Fig. 9. Transformed Kubelka functions [F(R)h]
1/2
for undoped and doped titania
samples to estimate the band gap energies.
Table 2
Band gap energies from UV–vis data on hydrothermal samples as compared with
Degussa P-25
Photocatalyst Band gap energy (eV)
TiO
2
Degussa P-25 2.9 (7)
TiO
2
3.0 (3)

sion degrees up to 30% at low initial phenol concentrations and up
to 15% at higher phenol concentrations were obtained. The higher
degradation degree and no detectable organic compounds for TFE
sample indicate the superiority of this catalyst. A little quantity of
organic compounds was identified after reactions in the presence
of TF and TE catalysts. Hydroquinone, p-benzoquinone and cate-
chol were detected as main reaction intermediates. The presence
of the organic intermediates was attributed to the reduced rate of
hydroxyl radical generation and phenol hydroxylation. The degra-
dation mechanism proposed is based on hydroxylated steps [29].
Full mineralization of organic compounds (as phenols and other
aromatics) on the surface of illuminated titania proceeds via many
steps, which make one-electron oxidation or reduction reactions
possible [30]. In the first step hydroxy radicals are generated by
the reaction between holes, resulted after photo-excitation of the
Fig. 10. Phenol conversion degree C
Ph
(%) after 5 hof UV illumination ( = 312 nm) for
the hydrothermally synthesized TiO
2
samples; 2 M and 0.2 M are the initial phenol
concentrations.
152 L. Diamandescu et al. / Materials Chemistry and Physics 112 (2008) 146–153
Fig. 11. Phenol conversion degree C
Ph
(%) under visible irradiation ( >380 nm) cat-
alyzed by Fe- and Eu-doped and codoped TiO
2
.
semiconductor, and the surface hydroxyl species of the catalyst.

In summary, the undoped titania gives the lowest phenol con-
version, and the transformation of phenol increases for the samples
with Fe or Eu, while the codoped Fe/Eu sample gives the highest
activity.
4. Conclusions
Iron- and europium-doped TiO
2
nanoparticles were obtained by
a hydrothermal route, at mild temperature and pressure (∼200
◦
C
and ∼15 atm, for 1 h). Rietveld refinements of the XRD patterns
reveal the exclusive presence of iron- and europium-doped anatase
phase in hydrothermally synthesized samples; the particle mean
size was less than 15 nm and the morphology was found to depend
on doping element. EXAFS analysis strongly support that both
Fe
3+
and Eu
3+
ions enter the TiO
2
lattice, by substituting the Ti
4+
ions. Ti
4+
replacement by the larger Fe
3+
and Eu
3+

and Eu
3+
in the TiO
2
host lat-
tice.
The photocatalytic activity of all hydrothermal samples in the
degradation reaction of phenol is much higher in the visible light
than in UV region. An important UV → visible absorption shift
(∼20 nm) has been evidenced for the sample TF. However, the best
photocatalytic activity in the photodegradation reaction of phe-
nol was evidenced for the hydrothermal sample TFE, in both UV
and visible light regions. The remarkable conversion degree of phe-
nol recommends the codoped specimen TFE, obtained by a simple
hydrothermal route at moderate temperature, as a promise visible
photocatalyst for the degradation of harmful organic compounds in
water. We hope that these results will stimulate further theoretical
and experimental works for a better understanding of mechanisms
and doping effects in photocatalytically active materials.
Acknowledgements
This work was supported by the Romanian Ministry of Educa-
tion and Research through Contract CEEX-Matnantech no. 23/2005.
We gratefully acknowledge the valuable assistance of Dr. Edmund
Welter and Dr. Dariusz Zajac (HASYLAB) during the EXAFS exper-
iments, as well as the whole scientific support of Prof. Eberhardt
Burkel and Dr. Radu Nicula (University of Rostock).
References
[1] M.V. Rao, K. Rajeshwar, V.R. Pal Verneker, J. DuBow, J. Phys. Chem. 84 (1980)
1987.
[2] L. Palmisano, M. Schiavello, A. Sclafani, C. Martin, I. Martin, V. Rives, Catal. Lett.

[21] J. Yu, G. Wang, B. Cheng, M. Zhou, Appl. Catal. B: Environ. 69 (2007) 171.
[22] N.N. Greenwood, T.C. Gibb, M
˝
ossbauer Spectroscopy, Champman and Hall Ltd.,
1971.
L. Diamandescu et al. / Materials Chemistry and Physics 112 (2008) 146–153 153
[23] D. Barb, L. Diamandescu, D. Tarabasanu, J. Phys. C6 (1976) 113.
[24] R.D. Shannon, Acta Crystallogr. 32 (1976) 751.
[25] S. Zhu, T. Shi, W. Liu, S. Wei, Y. Xie, C. Fan, Y. Li, Phys. B 396 (2007) 177.
[26] W. Li, A.I. Frenkel, J.C. Woicik, C. Ni, S. Ismat Shah, Phys. Rev. B 72 (2005)
155315–155321.
[27] Y. Miyake, H. Tada, J. Chem. Eng. Jpn. 37 (2004) 630.
[28] J. Tauc, R. Grigorovici, A. Vancu, Phys. Status Solidi 15 (1966) 627.
[29] S.I. Zabinsky, J.J. Rehr, A. Ankudinov, R.C. Albers, M.J. Eller, Phys. Rev. B 52 (1995)
2995.
[30] A. Sobczynski, Ł. Duczmal, W. Zmudzinski, J. Mol. Catal. A: Chem. 213 (2004)
225.