Available online at www.sciencedirect.

com

Acta Materialia 57 (2009) 5174–5185

www.elsevier.com/locate/actamat

Structural and optical properties of CaTiO3 perovskite-based

materials obtained by microwave-assisted hydrothermal synthesis:
An experimental and theoretical insight
Mario L. Moreira a, Elaine C. Paris a, Gabriela S. do Nascimento a, Valeria M. Longo e,
Julio R. Sambrano b, Valmor R. Mastelaro c, Maria I.B. Bernardi c, Juan Andrés d,
José A. Varela e, Elson Longo e,*
a
LIEC, Departamento de Quı́mica, UFSCar, Rod Washington Luiz, km 235, PO Box 676, CEP 13565-905 São Carlos, SP, Brazil
b
Grupo de Modelagem e Simulacßão Molecular, DM, UNESP, PO Box 473, 17033-360 Bauru, SP, Brazil
c
Instituto de Fı́sica de São Carlos, USP, PO Box 369, 13560-970 São Carlos, SP, Brazil
d
Departamento de Quı́mica Fı́sica y Analı́tica, Universitat Jaume I, 12071 Castello, Spain
e
LIEC, Instituto de Quı́mica, UNESP, PO Box 355, 14801-907 Araraquara, SP, Brazil

Received 27 February 2009; received in revised form 8 July 2009; accepted 13 July 2009
Available online 21 August 2009

Abstract

CaTiO3 powders were synthesized using both a polymeric precursor method (CTref) and a microwave-assisted hydrothermal
(CTHTMW) method in order to compare the chemical and physical properties of the perovskite-based material as a function of the syn-
thesis method. To this end, X-ray diffraction, Raman spectroscopy, inductively coupled plasma atomic emission spectroscopy and exper-
imental Ti and Ca K-edge X-ray absorption near-edge structure spectroscopy, as well as measurements of photoluminescence (PL)
emission, were used to characterize the typical bottom-up process of the CaTiO3 perovskite phase at different times. Detailed Rietveld
refinements show a random polycrystalline distortion in the powder structure, which can be associated with the tilting (a angle < O–Ti–
O) between adjacent TiO6 octahedra (intermediate range) for CTHTMW samples and an intrinsic TiO6 distortion (short range) in relation
to the polymeric precursor CTref sample. These properties were further investigated by first-principles calculations based on the density
functional theory at the B3LYP level. The relationship between this tilting on the PL profile is highlighted and discussed. Thus, a struc-
tural model derived from both experimental results and theoretical simulations reveals a close relationship between this tilting and the
presence of intermediate energy states within the band gap which are mainly responsible for PL emissions.
Ó 2009 Published by Elsevier Ltd. on behalf of Acta Materialia Inc.

Keywords: CaTiO3; Microwave-assisted hydrothermal; Rietveld; Photoluminescence; Ab initio calculation

1. Introduction electronic and optical properties. Therefore, physical princi-

Order–disorder effects are the keys to many unsolved
structural problems and unexplained structure-related prop-
erties in solid materials. In particular, structural order–dis-
order is always present in real materials and may play an
important role in technological applications by altering their
*
Corresponding author. Tel.: +55 16 3361 5215; fax: +55 16 3351 8350.
E-mail address: elson@iq.unesp.br (E. Longo).
ples that govern the structural state of a given perovskite and
how that state may change have long been the subject of
investigation and debate. Compounds with the perovskite
structure, ABO3, and its derivatives are perhaps the most
widely investigated owing to their significance to both fun-
damental research and the high potential for technological
applications because of their diverse physical properties
[1–5]. Their electronic structures and photophysical proper-
ties have been the goal of much research; in particular,

1359-6454/$36.00 Ó 2009 Published by Elsevier Ltd. on behalf of Acta Materialia Inc.

doi:10.1016/j.actamat.2009.07.019
 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185
several efforts have been devoted to studying the photolumi-
nescence (PL) emission of titanates such as ATiO3 (A = Ca,
Sr, Ba, Pb) [6–12].
The PL emission in perovskite materials is considered to
be closely related to the crystal structure and their corre-
sponding distorted metal–oxygen polyhedra. In particular,
TiO6 octahedra in this family exhibit marked structural
flexibility. Perhaps the most striking manifestation of such
flexibility is the fact that under ambient conditions many
titanates present structures with lower symmetries that
can indeed be derived from the cubic aristotype structure
(Pm3m space group) through the rotation or tilting of reg-
ular, rigid octahedra or due to the presence of distorted
TiO6 octahedra [13]. It is significant that although the
structure is distorted, the rotation/tilting does not disrupt
the corner-sharing connectivity.
The tilting of the octahedral framework also plays a major
role in determining the properties of oxide perovskites.
Cooperative rotations can give rise to as many as Glazer tilt
systems [13–16]. In the seminal studies of Glazer [13] and,
more recently, those of both Woodward [17,18], and
Howard and Stokes [16], the changes in symmetry resulting
from different types of octahedral tilting have been systema-
tized and rationalized. Glazer’s notation specifies the magni-
tude and phase of the rotation of the octahedra about the
three orthogonal axes of the aristotype cubic unit cell. For
example, in the CaTiO3 (CT) the notation is a b+a , which
implies unequal tilts about the x, y, and z axes. The positive
superscript indicates neighboring octahedra tilt in the same
direction (in phase) while a negative superscript indicates
the octahedra tilt in opposite directions (out of phase).
Although the precise mechanisms of these transformations
remain elusive, both experimental and theoretical studies
have shed light on these interesting phenomena.
The present work focuses on links between the structure
and the photoluminescent behavior of titanates with the
parent mineral CaTiO3 (CT) selected as a case study. In
the framework of a more extended project [19,20,21] aimed
at the design and synthesis of perovskite-based materials
with PL properties. We investigate the role of microwave
processing on the structural and photoluminescent behav-
ior of CT powders. The main objective is to investigate
the impact of structural order–disorder on the PL emission
associated to octahedral tilting. This research involves four
critical steps: (1) synthesis of powders; (2) structural disor-
der characterization to elucidate their relationship with the
parent ordered samples; (3) establishing the formation
mechanisms of the disorder; and more importantly, (4)
revealing the relationships between structural disorder
and the PL property. The purpose is to join both the expe-
rimental and theoretical results to explain the different
responses of PL emission at room temperature by using a
structural disorder motif. To this end, we have used differ-
ent experimental techniques and first-principle calculations
based on the density functional theory (DFT) to carry out
a comprehensive investigation of the corresponding struc-
tures and electronic properties.
5175
2. Experimental and theoretical procedures
2.1. Sample preparation
Perovskite-based materials can be synthesized by a
plethora of preparation methods. Mao et al. [22] describe
a number of advances that have been made in the synthesis
of various ABO3-type oxides. The most popular techniques
include the traditional solid state reaction method [23–26],
formation through a soft solution chemistry methodology
such as a co-precipitation method [27], a hydrothermal
method [28,29], a solverthermal method [30], alkoxide
hydrolysis [31], metal–organic processing [32] and Pechini
[33].
Hydrothermal media provide an effective reaction envi-
ronment for the synthesis of numerous ceramic materials
because of the combined effects of solvent, temperature,
and pressure on ionic reaction equilibrium. The conven-
tional hydrothermal (CH) method has become an effective
synthetic route for the materials science by dramatically
increasing control of the micro/nanometric morphology
and orientation [34]. In addition, this method is environ-
mentally friendly and depends on the solubility of the
chemical salts in water under temperature and pressure
conditions. The key factors in this method are the vapor
pressure and solubility of the chemical salts in water [35].
In contrast to the CH method which requires a long time
(typically half to several days) and high electric power (over
a thousand watts) [36], microwave-assisted heating is a
greener approach to synthesize materials in a shorter time
(several minutes to hours) and with lower energy consump-
tion (hundreds of watts). Consequently, the microwave-
assisted route is a rapidly developing area of research
[37–40] which has been demonstrated to be efficient in the
processing of many oxides [41,42]. Recently, the impact
of microwave frequency on the hydrothermal synthesis of
nanocrystalline tetragonal BaTiO3 has been analyzed by
Suib et al. [39], indicating that the HTMW process can
be a potential and breakthrough way for the synthesis of
crystalline CT powders at low temperatures and high heat-
ing rates.
The condition of Ca/Ti = 1 corresponds to the composi-
tion of the desired CaTiO3 product which was synthesized
using TiCl4 (99.99%, Aldrich), CaCl2
2H2O (99.9%,
Merck) and KOH (99%, Merck). Three solutions were pre-
pared: in the first solution 0.05 mol of the TiCl4 was slowly
added to 125 ml of de-ionized water at approximately 0 °C
under vigorous stirring, forming TiO(OH)2 + H+ + Cl .
Similarly, 0.05 mol of CaCl2
2H2O was dissolved in the
de-ionized water. Then, two precursor solutions containing
Ti4+ and Ca2+ ions were mixed, homogenized and shared
in five portions of 50 ml, to which 50 ml of the KOH solu-
tion (6 M) was added to act as a mineralizer [35] and bring
to pH = 14.
The suspension was loaded into a 110 ml Polytetraflu-
oroethene autoclave reaching 90% of its volume, providing
the maximum pressure efficiency to the system [43]. The
5176 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185
autoclave was sealed and placed into a microwave-assisted
hydrothermal system using 2.45 GHz microwave radiation
(Support information S1) with a maximum output power
of 800 W [44]. The reaction mixture was heated to 140 °C
in less than 1 min (at 800 W) by a direct interaction of
water with microwave radiation and was kept at this tem-
perature for 10 (CT10), 20 (CT20), 40 (CT40), 80 (CT80)
and 160 (CT160) min under pressure of 2.5 bar. Afterward
the autoclave was naturally cooled to room temperature.
Then the solid product was washed with de-ionized water
until a neutral pH was reached and then dried at 80 °C
for 12 h. The CaTiO3 reference sample (CTref) was pre-
pared by the polymeric precursor method [45,46] and
heat-treated at 700 °C for 2 h in a conventional furnace
under an air atmosphere and then naturally cooled to room
temperature. This sample did not present photolumines-
cence properties, and therefore the sample was used as a
reference sample (CTref).
2.2. Characterization techniques
Structural techniques were employed to provide infor-
mation on different time and length scales. Each level has
its particular complexity, and the results therefore offer
structural insight into those length scales that determine
many important aspects of material phenomenology and
their properties. Also, it is becoming increasingly clear that
a detailed microscopic structural understanding of materi-
als is necessary. Thus it is mandatory to use different tech-
niques such as X-ray diffraction, which enables an average
depiction of the structure (through pattern matching and
Rietveld analysis). Moreover, there is a wide range of
experimental techniques that can be considered as comple-
mentary tools to characterize specific properties of the
materials. Both Raman and UV spectroscopies are among
the advanced nondestructive characterization tools used to
acquire information on the estimate of the structural order
at short and medium ranges of a material. Therefore, they
can be used to estimate the crystal potential fluctuations
and local atomic arrangement. In addition, currently it is
well recognized that optical properties like PL depend on
both structural and electronic properties, including compo-
sitional ordering and the presence of impurities and defects.
PL depends on electronic excitations and thus is a neces-
sary complement to spectroscopies concerning lattice exci-
tations, yielding structural information of a completely
different character from the information obtained by dif-
fraction-based techniques. The latter detected long-range
order while the former yielded information on the immedi-
ate surroundings of an ion (particularly useful as a struc-
tural probe in short-range amorphous structures).
X-ray absorption near-edge structure (XANES) spectros-
copy provides information about the local order around
cations in oxide materials [47–50]. Ti K-edge XANES exper-
imental results of photoluminescent titanates revealed the
coexistence of two types of environment of titanium atoms,
namely fivefold coordinated TiO5 (square-base pyramid)
and sixfold coordinated TiO6 (octahedron), before reaching
the complete structural order [47,48]. Based on the Ca K-
edge L3,2 XANES spectra, Asokan et al. [49] reported
changes in the local order around Ca atoms.
Powder X-ray diffraction (XRD) experiments were car-
ried out on the as-prepared samples, and the corresponding
measurements were obtained by Rigaku DMax 2500PC
using Cu Ka1 (k = 1.5406 Å) and Cu Ka2 (k = 1. 54434 Å)
radiation. Data were collected from 20° to 120° in 2h range
with an 0.5° divergence slit and an 0.3 mm receiving slit
using a fixed-time mode with a 0.02° step size and 1s/point.
The Rietveld refinements [51] were carried out with GSAS
software [52], which is specially designed to simultaneously
refine both structural and microstructural parameters using
the least-square method. The peak profile function was
modeled using the convolution of the Thompson–Cox–
Hastings pseudo-Voigt (pV-TCH) with the asymmetry
function described by Finger et al. [53]. The background
of each pattern was fitted by a polynomial function.
Raman spectra were recorded on a RFS/100/S Bruker
Fourier transform Raman (FT-Raman) spectrometer with
a Nd:YAG laser providing an excitation light at 1064 nm
in a spectral resolution of 4 cm 1.
An inductively coupled plasma atomic emission spec-
trometer, ICP-AES Simultaneous CCD–VISTA–MPX
(Varian), with radial configuration, was used for chemical
analysis. The dissolution procedure was carried out using
10 ml of HCl (37% purity) and 3 ml of HNO3 (68% m/m)
in closed vessels for 2 h at room temperature. Analytical
blanks were prepared following the same acid digestion
procedure. Background signal correction was carried out
by the operating software of the instruments. The genera-
tor frequency was 40 MHz, with an RF power of 100 kW
and a plasma gas flow rate of 15 l min 1. Emission bands
at 396.8 and 366.1 nm were used to quantify calcium and
titanium content, respectively. Microstructural character-
ization was performed by field emission scanning electron
microscopy (FE-SEM, Zeiss SupraTM 35).
Titanium and calcium K-edge X-ray absorption spectra
were collected at the LNLS (National Synchrotron Light
Laboratory) facility using the D04B-XAS1 beam line.
The LNLS storage ring was operated at 1.36 GeV and
160 mA. XANES spectra of grounded samples were col-
lected at the Ti K-edge (4966 eV) and Ca K-edge
(4205 eV) in the transmission mode at room temperature
using a Si(1 1 1) channel-cut monochromator. XANES
spectra were recorded between 4910 and 5100 eV for Ti
and 4000 and 4200 eV for the Ca K-edge using energy steps
of 0.3 eV. For comparison purposes of the different sam-
ples, all spectra were background removed and normalized
using as unity the first EXAFS (extended X-ray absorption
fine structure) oscillation.
UV–visible absorption was recorded using the Cary 5G
spectrometer in total reflection mode by the integration
sphere. Photoluminescence (PL) spectra were collected with
a Thermal Jarrel-Ash Monospec 27 monochromator and a
Hamamatsu R446 photomultiplier. The 350.7 nm exciting
 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185 5177

Table 1 Yang–Parr gradient-corrected correlation functional,

Internal and lattice parameters used for theoretical calculations. B3LYP [56], which has proven to be a very effective tool to
CTref CT10 CT160 deal with the present challenging problem. We have been
Lattice parameters particularly successful in employing this functional to study
a (Å) 5.387 5.406 5.405 the electronic and structural properties of bulk and surfaces
b (Å) 5.439 5.492 5.489 of PbTiO3 [57,58], AZrO3 [7,59] and several other oxides
c (Å) 7.646 7.664 7.662
[60,61]. The atomic centers are described by all electron basis
Internal parameters sets 86-511(d21)G for Ca [62], 86-411(d31)G for Ti [62] and
Ti 0000, 0500, 0000 0000, 0500, 0000 0000, 0500, 0000
6-31G* for O (Support information S2). The XcrysDen pro-
Ca 0992, 0033, 0250 0989, 0044, 0250 0991, 0043, 0250
O1 0072, 0489, 0250 0085, 0484, 0250 0086, 0479, 0250 gram [63] was used for the band structure drawing design.
O2 0717, 0284, 0034 0706, 0292, 0037 0708, 0290, 0036 The analysis of the vibrational modes and their correspond-
CTref = reference sample, prepared by the polymeric precursor method. ing frequencies were calculated through numerical second
derivatives of the total energies as implemented in the
CRYSTAL06 package [64]. Lattice constants and internal
wavelength of a krypton ion laser (Coherent Innova) was coordinate data were obtained from XRD pattern refine-
used, with the nominal output power of the laser kept at ment using the Rietveld method to better describe the struc-
200 mW. All measurements were taken at room temperature. tural distortion derived from the experimental data.
CT crystallizes in an orthorhombic perovskite structure
2.3. Computational method and periodic model Pbnm space group, in a single phase with four nonequiva-
lent atoms per unit cell, with lattice parameters and inter-
The simulation was performed using a periodic approxi- nal coordinations obtained from refined parameters.
mation as implemented in the CRYSTAL06 computer code Internal coordinates and net parameters for CTref, CT10
[54]. Our computational method is the density functional and CT160 samples are listed in Table 1. The alpha (a)
theory in conjunction with Backe’s three parameter hybrid angle values for adjacent TiO6 octahedra are schematically
nonlocal exchange functional [55], combined with the Lee– illustrated in Fig. 1 together with their octahedral and

Fig. 1. Different values of a angles for (<O–Ti–O) adjacent [TiO6]–[TiO6] clusters and a schematic model for CTHTMW with octahedral and dodecahedral
sites for Ca and Ti atoms, respectively.
5178 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185

Table 2 3.2. X-ray diffraction and Rietveld refinements

Nominal and experimental compositions of CTHTMW samples.
Ceramic denomination Composition in molar fraction Diffraction patterns of the samples are depicted in
Nominal Ca/Ti Analyzeda Ca/Ti Fig. 2, featuring a single phase of polycrystalline CT. The
CT10 1 1.062 (±0.006) facility and low temperature synthesis of CT are attributed
CT20 1 1.065 (±0.002) to the fast reaction owing to the direct microwave coupling
CT40 1 1.055 (±0.003) with water rotational barrier, allowing a uniform solution
CT80 1 1.037 (±0.001) heating startup [65,66]. Both calcium and titanium hydrox-
CT160 1 1.023 (±0.003)
ide were previously described from a Pourbaix potential–
a
an = 5(a = analyzed, n = measures). pH equilibrium diagram [67] in aqueous solution at room
temperature. Lencka and Riman [35] suggested that signif-
icant amounts of mineralizers are necessary (i.e., pH adjust
agent) when non-alkaline precursors (absence of OH
group) are used. The direct rotational water excitation by
microwave radiation is able to uncouple OH groups from
calcium and titanium hydroxides, denuding the calcium
clusters and enabling them to interact with the titanium
hydroxide clusters much more readily than calcium clusters
themselves in aqueous media [66]. Therefore, under micro-
wave radiation heating and at a pH P 9, CT precipitation
[35] is favored as indicated by Scheme 1, enhancing the
kinetics crystallization of CT powders by one to two orders
of magnitude [38,41,42,68]. Thus, it may be presumed that
the mobility of calcium and titanium clusters is higher
under hydrothermal conditions than at ambient pressure
and temperature [69]. Effective collision rates occur when
particles collide, producing irreversible oriented attach-
Fig. 2. X-ray patterns of CT samples annealed at 140 °C from 10 to ments [70] which offer favorable thermodynamic and kinet-
160 min.
ics conditions for CT shaping. Therefore, the
aforementioned facts qualify the hydrothermal microwaves
dodecahedral sites built on the lattice parameters obtained method as a typical bottom-up process. The formation of
from Rietveld refinements. CT from aqueous calcium and titanium precursors can be
expressed by Scheme 1.
3. Results and discussion The polycrystalline nature of CT powders is expressed
by XRD patterns as shown in Fig. 2, which were identified
3.1. ICP analysis as an orthorhombic phase with Pbnm space group. The
results obtained from the Rietveld method (Support infor-
ICP-AES is capable of determining the Ca/Ti ratio as mation S3) using the ICSD cif number 74212 are depicted
well as the purity degree of sample. To estimate the accu- in Table 3. The refinement was continued up to a conver-
racy of the CT suspensions with lower salt concentrations, gence to be reached with the value of the quality factor
two series of 10 and 100 were diluted in de-ionized water (v2) approaching 1. In our case, great values of v2 for all
prepared from the original digested samples. Table 2 pre- patterns are reached as depict in Table 3.
sents nominal compositions of the CTHTMW compounds. The Rietveld refined parameters depicted in Table 3 pro-
Thus the current synthesis method can yield the high-purity vide small increases of a, b, and c lattice parameters for quasi-
CT powders with impurity concentrations of Fe, Sr and crystalline CTHTMW samples in relation to the crystalline
Mg ions less than 1 ppm. The results depicted in Table 2 CTref sample. Some modifications are observed in the values
are very important because tiny changes in the chemical of alpha angles (a) between adjacent [TiO6]–[TiO6] octahe-
composition induce many changes in chemical and physical dral clusters as shown in Table 3. and schematically reported
properties of the crystal structure. As secondary products in Fig. 1 while only small modifications were found in the
were removed by successive washes, all characteristics [CaO12] dodecahedral clusters. Therefore, the a value can
and properties observed for these samples will be relative be used as a borderline between a quasi-amorphous
only to CTHTMW compounds. (a  158°) to a quasi-crystalline (a < 158°) structure. The

CaCl 2 (aq ) + TiCl 4 (aq ) + 6 KOH (aq ) → CaTiO3 ( s ) + 6 KCl (aq ) + 3H 2 O(l )

Scheme 1. CaTiO3 formation reaction from aqueous calcium and titanium precursor salts.
 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185 5179

Table 3
Rietveld refined parameters of CTref and CTHTMW powders according to HTMW treatment time.
Ceramic composition a (Å) b (Å) c (Å) v2 R-Bragg (%) Rwp (%) a (degree) Cell volume (Å3)
CT10 5.406 5.492 7.664 1.14 39 11.64 154.28 227.59
CT20 5.408 5.493 7.666 1.44 27 6.87 154.31 227.54
CT40 5.407 5.492 7.665 1.25 47 11.60 154.04 227.64
CT80 5.406 5.491 7.664 1.28 40 12.01 154.46 227.54
CT160 5.405 5.489 7.662 1.22 44 11.47 155.22 227.40
CTref 5.387 5.439 7.646 1.22 38 10.35 158.23 224.09
v2, goodness of fit; CTref, reference sample, prepared by the polymeric precursor method; Rwp, weighted error (%); a, angle between adjacent [TiO6]–[TiO6]
octahedra.

value of the a angle for high ordered and non-photolumines- Raman active modes are commonly observed: 134 cm 1
cent orthorhombic CTref powders (approximately 158°) is for a vibration of Ca bonded to a TiO3 group (Ca–TiO3)
higher than the values of CTHTMW samples as pointed out lattice mode; 181, 224, 244, 287 and 339 cm 1 are associ-
in Table 3 for titanium octahedron [TiO6] tilts in the z direc- ated with O–Ti–O bending modes; 464 and 495 cm 1 corre-
tion (around y axis). This movement can actually be associ- spond to Ti–O6 torsional (bending or internal vibration of
ated with the existence of an orthorhombic–orthorhombic the oxygen cage) modes and finally the 669 cm 1 associated
structural transition, and this behavior can be related to with the Ti–O symmetric stretching mode [77–79]. Both
the tilt movement that dictates the orthorhombic symmetry experimental and theoretical values of Raman active
[16]. modes of CT powders are listed in Table 4 and depicted
in Fig. 4, together with well-defined peaks for the CTref
3.3. FE-SEM microscopy sample [80]. Duran et al. [81] suggested that the symmetry
of crystals observed by Raman spectroscopy indicates a
The micro-cube-shaped morphology of the CTHTMW local and dynamic symmetry while the XRD measurement
sample is quite similar to the morphology of analytical cal- is an average and static symmetry. From this point of view,
cite crystals [71] as shown in the FE-SEM image of Fig. 3. both methodologies confirm the orthorhombic structure of
In Fig. 3a the monodispersed micro-cube-shaped mor- CTHTMW samples.
phology with an edge length around 2 lm is clearly seen
which is different from the values obtained by the usual
synthesis methods [72], and in the current work it remains
practically unchanged. According to Wang [73], for face-
centered cubic (fcc) materials the multiple twinning of
small cubes occurs by the sharing of common crystallo-
graphic planes and then microstrains on CT cubes (see
Fig. 3b and c) can appear. The structure of one cube is
the mirror reflection of the other cube of the twin plane,
facilitating their adhesion and a possible crystal growth
by a coalescence process which emerges at extremely
extended times (much longer than 160 min). At this point,
it is important to remark that we can observe crystalline
nanoparticles with ordered superstructures on the scale of
several hundred nanometers to micrometers (Fig. 3b and
c) consisting of highly ordered polycrystalline building
units with scattering similar to a single crystal. This
description coincides with the term mesocrystal introduced
recently by Colfen and Antonietti [74–76].

3.4. Raman spectroscopy

Raman spectroscopy is a well-known and useful method

for investigating the behavior of symmetry changes in cera-
mic compounds. For the orthorhombic D2h 16 symmetry of
CT powder in a Pbnm space group, 117 vibrational modes
with four formula units per primitive cell (ZB = 4) are Fig. 3. FE-SEM microscopy of the CT samples processed using a
expected, but most of these modes cannot be detected microwave-assisted hydrothermal method for 10 min: (a) monodispersed
because of their low polarizabilities [77]. Hence, only nine micro-cubes; (b and c) emphasis on multi-faced CT cubes.
5180 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185

An analysis of the results depicted in Fig. 4 shows that in

the low-frequency range of 100–400 cm 1 the soft mode at
135 cm 1 is not commonly active for first-order Raman
scattering in a orthorhombic perovskite-type structure.
This experimental mode is seemingly due to a perturbation
of the perfect crystal symmetry by grain boundaries or sec-
ond order processes according to Zelezny et al. [82] while,
in the theoretical simulation, the mode can be related to a
periodic distortion on the orthorhombic structure which is
certainly related to an increase in the lattice parameters a,
b, and c. The bands centered at 181 and 224 cm 1 are asso-
ciated with the O–Ti–O bending mode. It is known that the
location of a morphotropic phase boundary depends on the
degree of orthorhombic lattice distortion, which is mainly
associated with the Ag band at 181 cm 1 that disappears
when phase transitions take place [78–83], indicating the
existence of a single orthorhombic phase for all situations.
Moreover, in the low-frequency range, the best defined and
strongest mode at 244 cm 1 is assigned to hard mode B1g as
a result of the rotation of the oxygen octahedron cage. The
modes at 287 and 339 cm 1 were dominated by Ti–O inte-
rior vibrations from the tilt at the titanium octahedron [84].
An analysis of Fig. 4 report that in the high frequency
Raman region at 400–800 cm 1, no drastic spectral change
is observed for the mode at 464 cm 1 as compared to CTref
and CTHTMW samples. On the other hand, for a torsional
mode at 495 cm 1, there is a displacement in an upward
direction rising to an intense and wide mode around
537 cm 1 which is interpreted as a signal of the disordered
structure by an increase in the tilts in the titanium octahe-
dron cluster as demonstrated by Hirata et al. [77]. From
this viewpoint, the uncommon upward displacement
toward the high-frequency region is expected to be affected
by an increase in the average electronic density. The extent
of distortion can be expressed by cell volumes which can to
be related to the average electronic densities at A-sites in
ABO3 compounds [85]. However, the ICP-AES results
from this work proved that only calcium occupies the A-
site. Even so, there is a significant volume increase in the
cell (Table 3) caused by the processing method which takes
the angle changes between the titanium octahedra and con-
Fig. 4. Raman shift for CTHTMW and CTref samples in orthorhombic
symmetry.
sequently distorts and polarizes the CTHTMW cell [77,83].
From Table 3 it can be observed that the superposition
of three Raman active modes at 518 (B2g), 530 (B1g) and
546 cm 1 (B3g) actually takes place, generating a wide band
centered at 537 cm 1. This vibrational mode was not found
in the CTref sample from quantum mechanical vibrational
simulations, just as it was not observed from experimental
spectrum (Fig. 4). Moreover, the very wide and low resolu-
tion band centered at 669 cm 1 for CTHTMW powders
shifts to upward frequencies in relation to the CTref sample
which is related to the Ag mode and described as a TiO6
distorted octahedron by Ti–O stretching. This fact is in
agreement with the already published data [77,78,86]. In
the theoretical simulations the absence of the vibrational
mode around 669 cm 1 can be attributed to low-resolution
modes at this region in conformity with Raman spectra of
CTHTMW samples which is in agreement with theoretical
and experimental results regarding the existence of the
TiO6 distorted octahedron for these samples.
3.5. XANES spectroscopy
It is well-known that perovskite-based materials (ABO3)
consist of network corner-linked BO6 octahedra enclosing
large cavities which form 12-coordination sites for the

Table 4
Theoretical and experimental Raman active modes (cm 1) for CaTiO3 samples.
Vibrational modes Ref. [21] Ref. [83] Ref. [85] CTref Exp CTref Theo CTHTMW Exp CTHTMW Theo
Ca–TiO3 128 – – 127 137 134 –
Lattice mode 160 155 153 156 157 – 157
O–Ti–O 181 180 178 180 200 181 181
Bending modes 220 226 222 224 227 224 224
245 247 244 245 240 244 237
295 286 281 288 285 287 257
339 337 333 339 362 339 315
Ti–O3 471 471 467 469 478 464 488
Torsional modes 495 495 490 495 499 537 518, 531, 546
Ti–O 644 639 – 645 – 669 –
Stretching mode
CTref, reference sample; ref, reference paper; Exp, experimental results; Theo, theoretical simulations.
 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185 5181

Fig. 5. The theoretical results for CT10 (HTMW): (a) CT160 (HTMW), (b) and CTref, (c) samples I–V report the band structure, total density of states,
the total and projected density states of axial oxygen atoms, the total and projected density states of equatorial oxygen atoms, and the projected density
states of titanium atoms, respectively.

A-site. In the present case, B- and A-sites correspond to Ti
and Ca elements, respectively, as represented in Fig. 1. PL
measurements induce energy transfer processes and specific
structural rearrangements occur; a remarkable evolution of
the distortion across the octahedral and dodecahedral clus-
ters can be sensed. Therefore, this fact may indicate a sym-
metry-breaking process along the network of both [BO6]
and [AO12] clusters, leading to a lower symmetry. Unavoid-
ably, one must consider the whole [BO6] and [AO12] poly-
hedra (i.e. both the octahedral and dodecahedral cluster
network) to study the overall pattern.
The luminescence property of titanates has been studied
by our group in recent years [87–91]. Studies show that,
from the analysis of XANES results, the quasi-amorphous
structures were associated with the presence of distorted
[TiO6]–[TiO5] clusters which are responsible for the lumi-
nescent emission. In addition, this behavior was related
to the organization of the sample, i.e. when the system pre-
sents a high level of crystallinity or is composed only of
ordered [TiO6]–[TiO6] clusters [92], the PL emission disap-
pears. In the present case, for CTref samples the Ti K-edge
XANES experimental results revealed the coexistence of
two types of environment for titanium atoms, namely, a
fivefold coordinated TiO5 (square-base pyramid) and a six-
fold coordinated TiO6, (octahedron) before reaching the
complete structural order. Fig. 6a displays the pre-edge
region of the XANES spectra of CTHTMW and CTref sam-
ples representing both situations, i.e. [TiO6]–[TiO5] (six-
fold–fivefold) and [TiO6]–[TiO6] (sixfold–sixfold). As can
be observed in Fig. 6a, the Ti K-edge spectra report a small
peak situated around 4971 eV, which is ascribed as 1s-3d
electronic transition [8,91]; this forbidden electronic transi-
tion can normally be allowed if the mixture of oxygen 2p
states and empty titanium 3d states [87] occurs. The
increase in the intensity of this peak indicates that the local
environment of Ti is noncentrosymmetric, distorting the
octahedral configuration [8,87,93]. XANES spectra remain
unchanged for all CTHTMW treatment times. However, the
intensity of the Ti K-edge pre-edge peak (Ti) is higher with
respect to the CTref crystalline sample and lower with
respect to the CTref disordered sample. These results point
out the existence of a noticeable disorder on still
unchanged TiO6 clusters of CTHTMW samples [94]. This
fact can be related to a quasi-crystalline structure by means
of tilted titanium octahedra in agreement with Rietveld
results Table 3. In other words, the Ti K-edge XANES
5182 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185

was achieved for all models and corresponds to 3.87, 3.86

Fig. 6. Ti (a) and Ca K-edge (b) XANES spectra of the CaTiO3 samples
processed in HTMW at 140 °C from 10 to 160 min and by the polymeric
precursor method (CTref).
and 3.98 eV for CT10, CT160 and CTref models, respec-
tively. The calculated direct band gap of the CTref model
agrees with the experimental optically measured gap.
Moreover, for both CTHTMW and for CTref samples
(Fig. 5), the upper valence band is mainly composed of
oxygen (O) 2p states, which are predominantly influenced
by changes on axial oxygen atoms (O9–O12), involved with
the alignment between two adjacent tilted TiO6 octahedra.
In the Ti octahedral site the axial oxygen corresponds to
atoms placed above and below the Ti–O plane in the octa-
hedral cluster Fig. 1 as a result of orthorhombic structure.
The O13–O20 represent de equatorial atoms Figs. 1 and 7
in Ti–O plane to orthorhombic structure by the CRYS-
TAL06 definition. The Ti (3d) contributions are clearly
predominant in the conduction band region. The structural
distortions existing in the titanium octahedron must
directly affect both the conduction and valence regions at
the same time.
An analysis of Fig. 5a and b points out the downward
slope of the energies for both valence and conduction
bands for CTHTMW in relation to the CTref system
Fig. 5c; the 0.1 eV reduction of the band gap value is
sensed. This behavior can be associated with the changes
of 2p and mainly 3d orbital redistributions in the valence
and conduction bands, respectively, as already indicated
by Ti K-edge spectra. UV–visible optical absorbance mea-
surements of CTHTMW (Support information S4) render

spectra present an unchanged local structure for CTHTMW

samples if the processing time remains from 10 min up to
160 min, indicating the same type of octahedral distortion.
The pre-edge region of the XANES spectra collected at
the Ca K-edge (Fig. 6b) shows no significant differences
between the CTHTMW samples compared to the CTref sam-
ples (pre-edge Ca). Thus, the energy transfer processes pre-
sented by the quasi-crystalline powders can be attributed to
the intrinsic disorder between [TiO6] and [CaO12] clusters,
originating a polarized orthorhombic CaTiO3 structure
(Fig. 1) without the presence of disordered [CaO12] and/
or [TiO5] clusters, which is in agreement with both Raman
spectra and theoretical data.

3.6. Theoretical modeling

The orbital distributions as well as the density of states

(DOS) are depicted in Fig. 5. To analyze the changes in the
electronic structure, it is beneficial to consider the band
structure, which can be compared independent of the crys-
tal space group. Figs. 5a-I, b-I and c-I present band struc-
tures for the CT10 (a), CT160 (b) and CTref (c) models,
Fig. 7. (a) Evolution of the PL profile for CaTiO3 samples annealed by 10
respectively. The top of the valence band (VB) is located up to 160 min under a 350.7 nm excitation wavelength at room temper-
at the C point in the same way that the bottom conduction ature, (b) their schematic representation of octahedral tilt with (c) wide
band is also located at the C point. Thus a direct band gap band model to illustrate the three steps of PL emissions.
 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185
band gap values of 3.53 and 3.60 eV for CT10 and CT160
quasi-crystalline structures, respectively. Thus this value is
0.3 eV smaller than the corresponding CTref (3.85 eV) crys-
talline sample, which is in agreement with both experimen-
tal and theoretical results.
Density of states (DOS) for titanium 3d orbitals in the
conduction band region are more significant and hence
more strongly affected than oxygen 2p orbitals in the same
region by the structural modifications. Unlike within the
valence band, the oxygen 2p orbitals are more affected by
structural distortion in the quasi-crystalline arrangement,
particularly the axial oxygen atoms (O9–O12) as depicted
in Figs. 5a(III), b(III), and c(III). These statements are sup-
ported by the observed changes among band structures in
Figs. 5a–c(I). Thus, the theoretical calculation agrees with
distortions around titanium clusters as previously
described by Rietveld refinements, which are associated
with the value of a angle and octahedral distortions.
3.7. Photoluminescence properties
Fig. 7 illustrates the PL evolutions of CTHTMW samples
synthesized by a hydrothermal microwave route under dif-
ferent times which can be considered as a trace for the for-
mation of CT mesocrystals. In the photoluminescent
response for the CTHTMW samples with a 350.7 nm excita-
tion source, it can be seen that the luminescence behavior is
composed of a broad luminescence band in the range of
350–580 nm with a more defined peak at 614 nm covering
a wide range of visible spectra. In addition, the profile is
typical of a multiphonon process, i.e. a system in which
relaxation occurs by several paths involving the participa-
tion of numerous states within the band gap of the material
[95] originating from by intrinsic defects of the material
owing to the absence of secondary phases as well as any
line of organic matter (recently reported for BaTiO3 also
prepared by the microwave-assisted hydrothermal method)
[19].
The wide band model illustrated in Fig. 7c provides the
three necessary steps for a PL emission to occur: (i) in the
first step, the excitation source, hm, corresponds to the
energy to promote a photon absorption from O 2p states
at the valence up to Ti 3d states inside the forbidden band
gap; (ii) after excitation, the recombination process occurs
among the excited 3d states closer to the conduction band;
and (iii) a wide band PL emission due to allowed 3d ? 2p
transitions, associated with a multiple hm‘ energy, can be
measured.
To discuss the PL profile of Fig. 7, the luminescence
spectra are broken up into five peaks, and each was fitted
to a symmetric Gaussian function (Support information
S5) Fig. 8 depicts the typical peak-resolution results, show-
ing that five Gaussian curves constitute the overall lumines-
cence in the visible region from approximately 380–580 nm.
These broad bands are composed of B1 and B2 (blue com-
ponents), followed by a G3 (green component), a Y4
(yellow component), and a O5 (orange component). In
5183
Fig. 8. Area percentage with respective errors bars for the PL evolution of
CaTiO3 samples annealed at 140 °C using a HTMW process.
addition, in Fig. 8 the contribution of each of the above
peaks as a function of time is presented. From the analysis
of these luminescence spectra, the following trends can be
derived.
(i) These transitions can be associated with intrinsic
defects in the CTHTMW quasi-crystalline (mesocrys-
tal) structure.
(ii) Both B2 and O5 present the large contributions.
(iii) The relative intensity of these bands tends to decrease
as the processing time increases except for the O5
peak in the orange region, which is now displaced
to a higher wavelength and a rise in the relative inten-
sity with increased processing time.
(iv) The contribution of the peaks measured by their cor-
responding area can be separated into three different
behaviors. First B1, G3 and Y4 present slight varia-
tions, B2 decreases from 42.8% to 33.5% while for
O5 an opposite behavior is sensed, rising from
19.2% to 38.9%.
(v) The wavelength values for B1, B2, G3 and Y4 can be
associated with electron–hole recombination centers
promoted by high energetic defects while O5 can be
related to low energetic defects [88].
(vi) The relative contribution of these transitions are
related to high energetic defects in the B1 peak, which
are not favored by an increase in the microwave
hydrothermal processing time; an opposite trend is
observed for the low energetic defects correlated with
the O5 peak.
Therefore, HTMW processing promotes an intrinsic dis-
order in a medium-range order in crystalline orthorhombic
powders by different angles concerning the TiO6 octahedra,
which are responsible for the PL property in the CTHTMW
quasi-crystalline samples. Thus, PL spectroscopy can be
used as a very sensitive probe to analyze the structural
order–disorder degree in the lattice of ceramic materials.
5184 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185
In several cases [96,97], it has been proposed that the Appendix A. Supplementary materials
intensity of the PL emission increases with a decrease in
the grain size, which is closely correlated with interfaces Supplementary data associated with this article can be
of nanocrystallites (<100 nm) and d-surface states inside found, in the online version, at doi:10.1016/j.actamat.2009.
the forbidden band gap by the distortion of TiO6 octahe- 07.019.
dra. However, in our case Fig. 3 the samples present a
micro-sized cube shape, and thus in our case the PL emis- References
sions cannot be attributed only to the size of the crystals
[1]
(nano-sized) as previously discussed. Then the origin of
[2]
the structure polarization and the consequent PL emission [3]
remain related to the structural distortion that is associated [4]
with TiO6–TiO6 tilting coupled with their internal
distortions. [5]
[6]
4. Conclusions
[7]
In summary, the present work reports on micro-cube-
shaped CaTiO3 powders prepared using HTMW methods. [8]
Their crystallization and optical properties were carefully
[9]
investigated and compared to powders synthesized by [10]
means of the polymeric precursor method. Using X-ray dif-
fraction, Raman spectra, ICP-AES spectroscopy and FE- [11]
SEM microscopy, it was verified that the HTMW process
[12]
is an efficient method to obtain CaTiO3 single-phase com-
pounds in a short processing time at low temperature. This [13]
method is highly cost-effective and environmentally [14]
friendly, qualifying this technique as a bottom-up process. [15]
Raman results combined with UV–visible absorbencies and [16]
the PL analysis demonstrate that structural disorder at a
[17]
short and intermediate range is present in CTHTMW pow- [18]
ders. On the basis of the TiO6 tilted octahedra and their [19]
connectivity (measured by the O–Ti–O angle between adja-
cent TiO6 clusters and the distorted octahedral site as sug- [20]
gested by Ti K-edge spectra), PL properties were
[21]
rationalized. The disorder coupled to the tilt of TiO6–
TiO6 adjacent octahedra was mainly responsible for the [22]
photoluminescent emission of a CTHTMW orthorhombic [23]
structure as a result of the remaining structural distortion. [24]
[25]
Theoretical results point out that these defects generate
localized electronic levels essentially above the valence [26]
band. This result was explained by the analysis of the band
structure and subsequent redistribution of the density of [27]
states mainly around titanium 3d and axial oxygen 2p
[28]
states. It is important to note that all distortions present
[29]
in the CTHTMW samples are dependent on the synthetic
route employed, a finding that needs further, more detailed [30]
evaluation. [31]
[32]
[33]
Acknowledgements
The authors acknowledge the support agencies CAPES, [34]
CNPq and FAPESP/CEPID 98/14324-8. J.A. acknowl-
edges to Ministerio de Educación y Cultura of the Spanish [35]
[36]
Government to provide research mobility. Thanks for
[37]
XANES facilities provided by LNLS, Campinas, SP, Bra- [38]
zil. Thanks to Rorivaldo Camargo and Madalena Tursi [39]
for the technical support.
Cohen RE. Nature 1992;358:136.
Scott JF. Ferroelectric memories. Berlin: Springer; 2000.
Dawber M, Rabe KM, Scott JF. Rev Mod Phys 2005;77:1083.
Kan DS, Terashima T, Kanda R, Masuno A, Tanaka K, Chu SC,
et al. Nat Mater 2005;4:816.
Maksimov EG, Matsko NL, Ebert SV, Magnitskaya MV. Ferroelec-
trics 2007;354:19.
Leite ER, Santos LPS, Carreno NLV, Longo E, Paskocimas CA,
Varela JA, et al. Appl Phys Lett 2001;78:2148.
Longo VM, de Figueiredo AT, de Lazaro S, Gurgel MF, Costa MGS,
Paiva-Santos CO, et al. J Appl Phys 2008;104:023515.
de Lazaro S, Milanez J, de Figueiredo AT, Longo VM, Mastelaro
VR, De Vicente FS, et al. Appl Phys Lett 2007;90:091906.
Leonelli R, Brebner JL. Phys Rev B 1986;33:8649.
Zhang WF, Yin Z, Zhang MS, Du ZL, Chen WC. J Phys Cond
Matter 1999;11:5655.
Orhan E, Varela JA, Zenatti A, Gurgel MFC, Pontes FM, Leite ER,
et al. Phys Rev B 2005;71:085113.
Lima RC, Espinosa JWM, Gurgel MFC, Paris EC, Leite ER, Joya
MR, et al. J Appl Phys 2006;100:034917.
Glazer AM. Acta Crystallogr, Sect B: Struct Sci 1972;B 28:3384.
Glazer AM. Acta Crystallogr A 1975;31:756.
Megaw HD, Darlington CNW. Acta Crystallogr A 1975;A 31:161.
Howard CJ, Stokes HT. Acta Crystallogr, Sect B: Struct Sci
1998;54:782.
Woodward PM. Acta Crystallogr, Sect B: Struct Sci 1997;53:32.
Woodward PM. Acta Crystallogr, Sect B: Struct Sci 1997;53:44.
Moreira ML, Mambrini GP, Volanti DP, Leite ER, Orlandi MO,
Pizani PS, et al. Chem Mater 2008;20:5381.
Moreira ML, Andrés J, Varela JA, Longo E. Cryst Growth Des
2009;9:833.
Marques VS, Cavalcante LS, Sczancoski JC, Volanti DP, Espinosa
JWM, Joya MR, et al. Solid State Sci 2008;10:1056.
Mao YB, Park TJ, Wong SS. Chem Commun 2005:5721.
Templeton LK, Pask JA. J Am Ceram Soc 1959;42:212.
Beauger A, Mutin JC, Niepce JC. J Mater Sci 1983;18:3041.
Hansen P, Hennings D, Schreinemacher H. J Am Ceram Soc
1998;81:1369.
Hennings DFK, Schreinemacher BS, Schreinemacher H. J Am Ceram
Soc 2001;84:2777.
Stockenhuber M, Mayer H, Lercher JA. J Am Ceram Soc
1993;76:1185.
Fukai K, Hidaka K, Aoki M, Abe K. Ceram Int 1990;16:285.
Wu MM, Long JB, Wang GG, Huang AH, Luo YJ, Feng SH, et al. J
Am Ceram Soc 1999;82:3254.
Chen DR, Jiao XL. J Am Ceram Soc 2000;83:2637.
Phule PP, Raghavan S, Risbud SH. J Am Ceram Soc 1987;70:C108.
Shaikh AS, Vest GM. J Am Ceram Soc 1986;69:682.
Pechini MP. Method of preparing lead and alkaline earth titanates
and niobates and coating method using the same to form a capacitor.
United States; 1967.
Luo ZJ, Li HM, Shu HM, Wang K, Xia JX, Yan YS. Mater Chem
Phys 2008;110:17.
Lencka MM, Riman RE. Chem Mater 1995;7:18.
Dutta PK, Asiaie R, Akbar SA, Zhu WD. Chem Mater 1994;6:1542.
Ma Y, Vileno E, Suib SL, Dutta PK. Chem Mater 1997;9:3023.
Sun WA, Li CH, Li JQ, Liu W. Mater Chem Phys 2006;97:481.
Nyutu EK, Chen CH, Dutta PK, Suib SL. J Phys Chem C
2008;112:9659.
 M.L. Moreira et al. / Acta Materialia 57 (2009) 5174–5185
[40] Dorner-Reisel A, Schops S, Lenk A, Schmutzler G. Adv Eng Mater
2007;9:400.
[41] Komarneni S, Roy R, Li QH. Mater Res Bull 1992;27:1393.
[42] Volanti DP, Keyson D, Cavalcante LS, Simoes AZ, Joya MR, Longo
E, et al. J Alloys Compd 2008;459:537.
[43] Walton RI. Chem Soc Rev 2002;31:230.
[44] Volanti DP, Cavalcante LS, Keyson D, Lima RC, De Moura AP,
Moreira ML, et al. Met Mater 2007;63:352.
[45] de Figueiredo AT, Longo VM, de Lazaro S, Mastelaro VR, De
Vicente FS, Hernandes AC, et al. J Lumin 2007;126:403.
[46] Kakihana M. J Sol-Gel Sci Technol 1996;6:7.
[47] Jan JC, Kumar KPK, Chiou JW, Tsai HM, Shih HL, Hsueh HC,
et al. Appl Phys Lett 2003;83:3311.
[48] Pontes FM, Longo E, Leite ER, Lee EJH, Varela JA, Pizani PS, et al.
Mater Chem Phys 2003;77:598.
[49] Asokan K, Jan JC, Chiou JW, Pong WF, Tsai MH, Chang YK, et al.
Solid State Chem 2004;177:2639.
[50] Montoncello F, Carotta MC, Cavicchi B, Ferroni M, Giberti A,
Guidi V, et al. J Appl Phys 2003;94:1501.
[51] Rietveld HM. J Appl Crystallogr 1969;2:65.
[52] A.C. Larson RBVD. Los Alamos National Laboratory Report, vol.
No. LAUR 86-748, 2004.
[53] Finger LW, Cox DE, Jephcoat AP. J Appl Crystallogr 1994;27:892.
[54] Dovesi VRS, Roetti C, Orlando R, Zicovich-Wilson CM, Pascale F,
Civalleri B, et al. CRYSTAL06 user’s manual. Torino: University of
Torino; 2006.
[55] Becke AD. J Chem Phys 1993;98:5648.
[56] Lee CT, Yang WT, Parr RG. Phys Rev B 1988;37:785.
[57] de Lazaro S, Longo E, Sambrano JR, Beltran A. Surf Sci
2004;552:149.
[58] Alicea J, Fisher MPA. Phys Rev B 2007;75:144411.
[59] Cavalcante LS, Longo VM, Zampieri M, Espinosa JWM, Pizani PS,
Sambrano JR, et al. J Appl Phys 2008;103:063527.
[60] Xu Y, Xiong GG. Chem Phys Lett 2004;388:330.
[61] Sambrano JR, Orhan E, Gurgel MFC, Campos AB, Goes MS, Paiva-
Santos CO, et al. Chem Phys Lett 2005;402:491.
[62] http://www.crystal.unito.it/Basis$_{-}$Sets/Ptable.html.
[63] Kokalj A. J Mol Graphics Modell 1999;17:176.
[64] Pascale F, Zicovich-Wilson CM, Gejo FL, Civalleri B, Orlando R,
Dovesi R. J Comput Chem 2004;25:888.
[65] Wilson GJ, Matijasevich AS, Mitchell DRG, Schulz JC, Will GD.
Langmuir 2006;22:2016.
[66] Rao KJ, Vaidhyanathan B, Ganguli M, Ramakrishnan PA. Chem
Mater 1999;11:882.
[67] Pourbaix M. Atlas of electrochemical equilibria in aqueous solu-
tions. Houston: NACE International Cebelcor; 1974.
[68] Komarneni S, Rajha RK, Katsuki H. Mater Chem Phys 1999;61:50.
[69] Rabeneau A. Angew Chem Int Ed Engl 1985;24:1026.
5185
[70] Godinho M, Ribeiro C, Longo E, Leite ER. Cryst Growth Des
2008;8:384.
[71] Matsubara H, Hayashi T, Ogawa T, Muramoto K, Jimbo M, Kamiya
H. Fish Sci 2008;74:418.
[72] Alfredsson M, Cora F, Dobson DP, Davy J, Brodholt JP, Parker SC,
et al. Surf Sci 2007;601:4793.
[73] Wang ZL. J Phys Chem B 2000;104:1153.
[74] Colfen H, Antonietti M. Angew Chem Int Ed Engl 2005;44:5576.
[75] Meldrum FC, Colfen H. Chem Rev 2008;108:4332.
[76] Colfen H, Antonietti M. Mesocrystals and nonclassical crystalliza-
tion. Chichester: John Wiley; 2008.
[77] Hirata T, Ishioka K, Kitajima M. J Solid State Chem 1996;124:353.
[78] Zheng H, Reaney IM, de Gyorgyfalva G, Ubic R, Yarwood J, Seabra
MP, et al. J Mater Res 2004;19:488.
[79] Malibert C, Dkhil B, Kiat JM, Durand D, Berar JF, SpasojevicdeBire
A. J Phys Condens Matter 1997;9:7485.
[80] Pontes FM, Escote MT, Escudeiro CC, Leite ER, Longo E, Chiquito
AJ, et al. J Appl Phys 2004;96:4386.
[81] Duran P, Gutierrez D, Tartaj J, Banares MA, Moure C. J European
Ceram Soc 2002;22:797.
[82] Zelezny V, Cockayne E, Petzelt J, Limonov MF, Usvyat DE,
Lemanov VV, et al. Phys Rev B 2002:66.
[83] Li Y, Qin S, Seifert F. J Solid State Chem 2007;180:824.
[84] Lukeman PS, Stevenson ML, Seeman NC. Cryst Growth Des
2008;8:1200.
[85] Milman V, Winkler B. Int J Inorg Mater 1999;1:273.
[86] Balachandran U, Eror NG. Solid State Commun 1982;44:815.
[87] Volanti DP, Cavalcante LS, Paris EC, Simoes AZ, Keyson D, Longo
VM, et al. Appl Phys Lett 2007:90.
[88] Longo VM, Cavalcante LS, Erlo R, Mastelaro VR, de Figueiredo
AT, Sambrano JR, et al. Acta Mater 2008;56:2191.
[89] Pizani PS, Leite ER, Pontes FM, Paris EC, Rangel JH, Lee EJH,
et al. Appl Phys Lett 2000;77:824.
[90] Paris EC, Espinosa JWM, de Lazaro S, Lima RC, Joya MR, Pizani
PS, et al. Chem Phys 2007;335:7.
[91] Moreira ML, Gurgel MFC, Mambrini GP, Leite ER, Pizani PS,
Varela JA, et al. J Phys Chem A 2008;112:8938.
[92] de Figueiredo AT, de Lazaro S, Longo E, Paris EC, Varela JA, Joya
MR, et al. Chem Mater 2006;18:2904.
[93] Vedrinskii RV. J Phys Condens Matter 1998;10:9561.
[94] Farges F, Brown GE, Rehr JJ. Phys Rev B 1997;56:1809.
[95] Anicete-Santos M, Cavalcante LS, Orhan E, Paris EC, Simoes LGP,
Joya MR, et al. Chem Phys 2005;316:260.
[96] Gu HS, Hu ZL, Hu YM, Yuan Y, You J, Zou WD. Colloid Surf A –
Physicochem Eng Asp 2008;315:294.
[97] Meng JF, Huang YB, Zhang WF, Du ZL, Zhu ZQ, Zou GT. Phys
Lett A 1995;205:72.