Study of the process of mechanochemical activation to obtain Aurivillius oxides with n=1

Article abstract of DOI:10.1006/jssc.2001.9188

Mechanochemical activation has been successfully used as an alternative method for producing Aurivillius oxides with n=1, like Bi₂MoO₆, Bi₂VO_5.5, and Bi₂VO₂. The analysis of the mechanoacti

Full text of DOI:10.1006/jssc.2001.9188

Journal of Solid State Chemistry 160, 54}61 (2001)
doi:10.1006/jssc.2001.9188, available online at http://www.idealibrary.com on
Study of the Process of Mechanochemical Activation to Obtain
Aurivillius Oxides with n ؍ 1
J. Ricote, L. Pardo, A. Castro, and P. Millan
Instituto de Ciencia de Materiales de Madrid, CSIC, Cantoblanco, 28049 Madrid, Spain
Received December 22, 2000; in revised form March 13, 2000; accepted March 26, 2001; published online May 30, 2001
although this compound and its polymorphic phases are
Mechanochemical activation has been successfully used as an
alternative method for producing Aurivillius oxides with n ؍ 1,
like Bi2MoO6, Bi2VO5.5, and Bi2VO2. The analysis of the
mechanoactivated materials by SEM and TEM allows the study
of the mechanisms involved in the process of mechanical activa-
tion and the occurrence of mechanosynthesis. It is observed that
if the amorphization of all the starting oxides is achieved during
the process, evidence of mechanosynthesis, i.e., the synthesis of
the 5nal product, is found in the resulting particles. The phases
obtained either cannot be synthezised by the traditional solid-
state reactions or show special characteristics, all of which con-
more commonly used as oxidation catalysts (6). The case of
bismuth}vanadium oxides is more complex, because al-
though ferroelectricity has been studied (4, 7) and used for
applications in gas sensors, thermistors, or positive temper-
ature coe$cient resistors (PTCR) (8), their promising fer-
roelectric properties are restricted in applications due to
their high dielectric losses. Most of the e!orts have been
invested in improving the properties of Bi VO oxide by
ꢀ
ꢄꢅꢄ
doping (9), while only a few reports on the ceramic process-
ing and characterization of the equally interesting Bi VO
ꢀ
ꢄ
5rms that the mechanisms of phase formation involved in the composition are found in the literature (10, 11). In this paper
mechanical and the thermal activations are di4erent. Further
annealing at higher temperatures makes all these structures go
back to the phases traditionally obtained by thermal activa-
tion. ( 2001 Academic Press
Key Words: mechanochemical activation; Aurivillius; SEM;
TEM.
we study the problem of both compositions. A further point
of interest of these vanadium}bismuth oxides resides in the
fact that they exhibit three polymorphic phases (a, b, and c)
which are obtained at di!erent temperatures, "rst observed
in Bi VO (12) and later on Bi VO (13). The high con-
ductivity of the c phase, which appears at 5623C, has attrac-
ted a large amount of interest because of its applications in
solid-state fuel cells (14) or for lithium rechargeable batteries
(15). In this way, a wide range of substitutional solid solu-
ꢀ
ꢄꢅꢄ
ꢀ
ꢄ
1. INTRODUCTION
tions based on Bi VO
has been prepared in order to
ꢀ
ꢄꢅꢄ
The family of bismuth-containing compounds with stabilize the c phase at room temperature. This family of
layered perovskite structures, called Aurivillius oxides after compounds is known as BIMEVOX (16, 17). However, par-
the researcher who "rst synthesized and studied them (1), tial substitution of vanadium leads to a degradation of the
has received a large amount of attention in recent years due physical properties, i.e., the ionic conductivity.
to their interesting structural and physical properties. Crys-
Therefore, the exploration of alternative synthesis routes
tal structures consist of Bi O layers interleaved with n for this family of compounds seems to be necessary in order
ꢀ ꢀ
L>ꢁ L ꢂL>ꢁ
perovskite layers of A B O
, where A is a metal ion in to overcome the problems related to the traditional
12 coordination (like Sr, Ba, Bi) and B another metal ion in methods in use, like high processing temperatures and long
6 coordination (like Ti, Nb, Ta, V, Mo). The simplest case reaction times, as well as to improve the physical properties
corresponds to those members of the family with n"1, i.e., without the need of additives or dopants. Mechanochemical
only one layer of BO octahedra between bismuth-oxygen activation of inorganic compound mixtures has been tradi-
ꢃ
layers. Among them we "nd Bi MoO (2) and the oxygen- tionally applied to obtain nanocrystals, intermetallic com-
ꢀ
ꢃ
de"cient structures Bi VO and Bi VO (3, 4). Regarding pounds, or amorphous materials from crystalline phases,
ꢀ
ꢄ
ꢀ
ꢄꢅꢄ
the technological interest of Aurivillius compounds with whose mechanisms seem to be di!erent to usual ther-
n"1, we should mention that a large number of them are mochemical activation (18}20). The mechanically activated
ferroelectric polar materials and some show good properties materials become in most cases more reactive and capable
as ionic conductors. The occurrence of ferroelectricity of faster reactions and have been used either as an alterna-
in bismuth molybdenum oxides was "rst reported in (5), tive preparation route (21) or for the synthesis of novel
54
0022-4596/01 $35.00
Copyright ( 2001 by Academic Press
All rights of reproduction in any form reserved.

STUDY OF THE PROCESS OF MECHANOCHEMICAL ACTIVATION
55
materials (22). Mechanochemical activation has been suc-
cessfully used to synthesize the oxides analyzed in this work,
Bi VO , Bi VO (23, 24), and Bi MoO (25), as well as 3.1. X-ray Powder Diwraction
3. EXPERIMENTAL RESULTS
ꢀ
ꢄꢅꢄ
ꢀ
ꢄ
ꢀ
ꢃ
other complex Aurivillius oxides (26). It has been shown
that an energetic mechanochemical activation in a vibrating
mill of oxide mixtures, followed by annealing at moderate
temperatures, results in new phases stable at room temper-
ature like the highly conducting c-phase and the non-
previously reported #uorite phases of Bi VO and
Starting oxides are milled until the mixture appears to be
amorphous, a process that is monitored by XRD analysis, as
shown in Fig. 1. The initial pattern contains peaks coming
from bismuth, vanadium, and molybdenum oxides, those
from Bi O being particularly strong. They disappear after
ꢀ ꢂ
ꢀ
ꢄ
milling times of 72 h for the 2Bi O :V O mixture (Fig. 1a)
ꢀ ꢂ ꢀ ꢄ
ꢀ
Bi MoO (24, 25).
ꢀ
ꢃ
and 168 h for the Bi O :VO mixture (Fig. 1b) and
ꢀ ꢂ
So far, X-ray di!raction (XRD) analysis of the powder
was used to characterize the resulting products of the
mechanoactivated particles. However, this macroscopic
technique is not appropriate for the study of their composi-
tion and local structure in order to understand the e!ects of
the activation process and the subsequent thermal evolution
of the powder in the ceramic fabrication. We have chosen
for this study the simplest case, n"1 Aurivillius oxides, and
among them, three structures with di!erent degree of oxy-
gen occupancy: Bi MoO , Bi VO , and Bi VO One of
Bi O :MoO (Fig. 1c). The patterns lacking of di!raction
ꢀ ꢂ
ꢂ
peaks could correspond in principle to a collection of
amorphous particles, although the existence of small quant-
ities of nanocrystals will be undetectable by X-ray di!rac-
tion due to line broadening. To clarify this point and to
identify the di!erent phases resulting from the mechanoac-
tivation process, microscopic characterization techniques
are used, and the results are reported in the next section.
ꢀ
ꢃ
ꢀ
ꢄꢅꢄ
ꢀ
ꢄꢅ
the questions approached in this work is whether the syn-
thesized material is already present in the activated powder,
i.e., if a proper mechanosynthesis has taken place, and, if so,
to what extend and which is the resulting composition.
3.2. Electron Microscopy, Electron Diwraction,
and Compositional Analysis
3.2.1. Bi <O
ꢄꢅꢄ
ꢀ ꢄ
oxide. SEM studies of the 2Bi O :
ꢀ
ꢀ ꢂ
Microscopic characterization techniques, like transmission V O oxide mixture (Fig. 2) reveal how the mechanochemi-
electron microscopy (TEM) and energy dispersive X-ray cal activation produces a large decrease in the particle size
spectroscopy (EDS), have been used in these oxides with of the starting oxide particles (Fig. 2a), leading to submic-
n"1 Aurivillius structure.
rometer particles that are almost spherical in shape
(Fig. 2b). The mechanoactivated powder does not show
particles with di!erent features, which is an indication that
a homogenization process of the mixture has taken place.
2. EXPERIMENTAL PROCEDURE
About 2 g of stoichiometric mixtures of analytical grade After annealing this powder at 3853C, the c-Bi VO phase
ꢄꢅꢄ
Bi O and V O , VO , or MoO were homogenized by is identi"ed by XRD (24). The SEM image of the annealed
ꢀ
ꢀ ꢂ
ꢀ ꢄ
ꢀ
ꢂ
hand in an agate mortar and then placed in an agate or powder (Fig. 2c) shows a smaller particle size than the milled
stainless-steel pot along with a 5-cm diameter ball. The mixture. This fact re#ects particle shrinkage following crys-
sample was mechanochemically activated in a vibrating mill tallization, and shows that no signi"cant crystal growth has
(Fritsch Pulverisette 0) in air, for times ranging from 1 h to occurred in this system, indicating that the thermal treat-
3 weeks, as reported elsewhere (24, 25). Process was ment induces mainly the crystallization of the
monitored by XRD. Milled powders of 2Bi O :V O and mechanochemically activated powder.
ꢀ ꢂ ꢀ ꢄ
Bi O :MoO mixtures are annealed at moderate temper-
TEM studies reveal the presence of a majority of amorph-
ꢀ ꢂ
ꢂ
atures in air, while for the Bi O :VO mixtures we use an ous particles in the mechanically activated powder. How-
atmosphere of N to avoid oxidation. Dispersed particles of ever, some crystalline particles have also been observed. The
the milled and annealed powders were characterized by amorphous particles show sizes similar to the ones observed
ꢀ ꢂ
ꢀ
ꢀ
scanning and transmission electron microscopy. In order to by SEM (large particle in Fig. 3) and crystalline particles are
separate agglomerates and disperse the particles, slight grin- around 100}200 nm (crystal marked with an arrow in
ding of the powder in an agate mortar is followed by Fig. 3). The small size of these crystallites cause a line
ultrasonically vibration of the suspension until it is uniform. broadening e!ect in the X-ray di!raction pattern, which,
A drop of this suspension is placed onto a platinized Si together with the small quantities observed, makes them
wafer, dried out, and covered with a very thin layer of undetectable (Fig. 1a). EDS analysis of a large amount of
sputtered gold for SEM observation. In the case of TEM, crystalline particles shows the presence of Bi only, which
a Formvar-covered 200-mesh Cu grid with evaporated C is corresponds to one of the starting oxides, a-Bi O , as identi-
ꢀ ꢂ
used. Experiments were carried out with an ISI-DS-130C "ed by electron di!raction. Some other particles show Bi:V
SEM, working at 18 kV, and with a JEOL-2000FX TEM ratios of 2:1 or 1:1. This may correspond to crystallites with
working at 200 kV and equipped with EDS.
the desired composition, Bi VO , and others of a phase
ꢀ
ꢄꢅꢄ

56
RICOTE ET AL.
"ed as an intermediate phase in the formation of Bi VO
from the starting oxide mixture. The di!raction patterns
from the crystals with the desired "nal compositions (inset
ꢀ
ꢄꢅꢄ
in Fig. 3) can be indexed as the c-Bi VO
phase. In
ꢀ
ꢄꢅꢄ
agreement with the structural features of the c phase, no
superlattice re#ections are observed. The amorphous par-
ticles analyzed by EDS contain Bi:V ratios ranging from 1:1
to 1.5:1. The fact that this ratio is below the desired 2:1 may
FIG. 1. Evolution of the XRD patterns of the oxide mixtures with the
milling time: (a) 2Bi O :V O , (b) Bi O :VO ., (c) Bi O :MoO .
ꢀ
ꢂ
ꢀ
ꢄ
ꢀ
ꢂ
ꢀ
ꢀ
ꢂ
ꢂ
that may correspond to BiVO , which is also possible for
the same oxidation state of vanadium. This has also been
FIG. 2. SEM micrographs of particles from the 2Bi O :V O mixture:
(a) starting oxides, (b) mechanochemical activated mixture, (c) after anneal-
ꢆ
ꢀ
ꢂ
ꢀ ꢄ
observed previously in studies of this system (23) and identi- ing at 3853C.

STUDY OF THE PROCESS OF MECHANOCHEMICAL ACTIVATION
57
As only a few particles in the milled powder appear to
have been crystallized, the annealing process of this powder
will mainly produce the crystallization of the material. This
is corroborated by the shrinkage of the particles observed
by SEM and also by previously reported DTA results (24).
They show a narrow exothermic peak at 3853C, the anneal-
ing temperature, which is usually related to a crystallization
process of an already synthesized product.
3.2.2. Bi <O oxide. The mechanoactivated powder of
ꢀ
ꢄ
the Bi O :VO oxides mixture was also studied by SEM
ꢀ ꢂ
ꢀ
(Fig. 4). Similar to the case of 2Bi O :V O mixture,
ꢀ ꢂ ꢀ ꢄ
mechanochemical activation produces a large decrease in
the particle size, leading to submicrometer particles that are
almost spherical in shape (Figs. 4a and 4b). Again, the
absence of two di!erent types of particles indicates a process
of homogenization of the mixture by mechanical activation.
A thermal treatment at 2753C of the milled mixture results
in agglomerates of 3}5 lm size mixed with some submic-
rometer particles similar to the ones from the milled powder
(Fig. 4c). This fact suggests "rst the sintering of the agglom-
erates of particles during the annealing and, second, that
this process has not been completed, as the presence of
submicronmeter particles suggests.
TEM studies reveal the presence of both crystalline and
amorphous particles in the initial, mechanically activated
powder. Crystalline particle sizes are around 100 nm
(Fig. 5a). Their size, and, again, their relative small quantity
within the powder, made them undetectable by normal
XRD analysis for the same reasons exposed in the previous
case. Amorphous particles are found to be larger, around
500 nm like the particle shown in Fig. 5b. EDS analysis of
FIG. 3. TEM micrograph of several particles from the mechanoac-
tivated 2Bi O :V O mixture. The large particle is identi"ed as amorph-
ous by electron di!raction. The smaller particle, marked with an arrow, is
ꢀ
ꢂ
ꢀ ꢄ
crystalline of the c-Bi VO
di!raction pattern.
phase. (Inset) The corresponding electron
ꢀ
ꢄꢅꢄ
indicate the presence in the amorphous particles of a certain the crystalline particles reveals Bi:V ratios far from the
percentage of BiVO (Bi:V"1:1). Sometimes the &&halo desired 2:1, most of them being characteristic of particles of
ꢆ
rings,'' characteristic of amorphous material, appear more VO , although some are identi"ed as the other starting
ꢀ
de"ned, probably indicating an incipient crystallization of oxide Bi O (Fig. 5a). Very rarely a crystalline particle with
ꢀ ꢂ
these particles.
the Bi:V ratio 2:1 of Bi VO is found. Regarding the
ꢀ
ꢄ
The electron di!raction analysis of the annealed powder amorphous particles analyzed, they also present, when ana-
reveals the presence of crystals of the intermediate BiVO lyzed by EDS, Bi:V ratios that suggest that they are mostly
phase together with crystals of the desired Bi VO . Due to Bi O particles.
ꢆ
ꢀ
ꢄꢅꢄ ꢀ ꢂ
their small size (100}200 nm, see Fig. 2b) and tendency to The mechanical activation has produced in this case
agglomerate, both types of crystals are di$cult to separate. amorphous particles mainly of one of the starting products
Nevertheless, EDS analysis of a group of such crystals Bi O . The e!ect of the mechanical activation on the crys-
ꢀ ꢂ
shows again intermediate values between 1:1 and 1.5:1 of tals of VO is restricted to a decrease in the crystallite size,
ꢀ
the Bi:V ratio, the same values obtained for the amorphous without any signi"cant amorphization. This fact may con-
particles. This is again probably resulting from an averaging tribute to the almost nonexistent production at this stage of
between the two types of crystals with ratios 1:1 and 2:1.
synthesized particles with the "nal composition (Bi VO ). If
ꢀ
ꢄ
According to these results, mechanical activation has we consider that amorphous particles are highly reactive,
produced amorphous particles of Bi VO that may also the absence of amorphous particles from one of the starting
ꢀ
ꢄꢅꢄ
contain a certain amount of synthesized BiVO . Crystallites oxides may prevent mechanosynthesis from taking place.
ꢆ
of these compositions have been also observed. Therefore, in All this can be also related to previous DTA results (24),
this case we can a$rm that a process of mechanosynthesis showing a wide peak at 2753C, which suggests that the
has taken place as a consequence of the mechanical activa- synthesis of the "nal product, and not only its crystalliza-
tion of the oxides mixture.
tion, takes place when the mechanically activated powder is

58
RICOTE ET AL.
During the TEM observations, and due to the exposure
to the electron beam, i.e., the supply of energy to the system,
changes in the particles of the annealed powder are
FIG. 4. SEM micrographs of particles from the Bi O :VO mixture:
ꢀ
ꢂ
ꢀ
(a) starting oxides, (b) mechanochemical activated mixture, (c) after anneal-
ing at 2753C.
heated up to that temperature. Therefore, annealed particles
tend to agglomerate (Fig. 4), most probably due to the
interdi!usion among the oxide particles that is taken place
at this stage. As we have already mentioned, there are
indications that this transformation is not complete during
annealing. The evolution of the DTA curve shows the occur-
FIG. 5. TEM micrographs of particles from the mechanoactivated
Bi O :VO mixture. The insets show the corresponding electron di!rac-
ꢀ
ꢂ
ꢀ
rence of further processes when increasing the temperature. tion patterns. (a) a-Bi O crystalline particle, (b) amorphous particle.
ꢀ
ꢂ

STUDY OF THE PROCESS OF MECHANOCHEMICAL ACTIVATION
59
observed in situ. The small particles collapse and produce
larger ones. Since these crystals have not "nished the pro-
cess of formation, they are not stable and any energy input
in the system results in further transformations.
3.2.3. Bi MoO oxide. SEM micrographs showing dif-
ꢀ
ꢃ
ferent stages of the Bi O :MoO oxide mixture are shown
ꢀ ꢂ
ꢂ
in Fig. 6. As before, mechanochemical activation produces
FIG. 7. TEM micrograph of particles from the mechanoactivated
Bi O :MoO mixture. (Inset) The electron di!raction spots obtained from
ꢀ
ꢂ
ꢂ
the particle marked with an arrow.
a large decrease in the particle size and an homogenization
of the starting oxide particles (Figs. 6a and 6b). When the
mechanically activated powder is annealed at 3003C, we
obtain the #uorite phase of Bi MoO , identi"ed by XRD in
ꢀ
ꢃ
(25), which is accompanied with a shrinkage of the particles
(Fig. 6c). As stated in the 2Bi O :V O system, which shows
ꢀ ꢂ ꢀ ꢄ
similar features (Fig 2c), this indicates that annealing indu-
ces mainly crystallization of the mechanically activated
powder.
TEM studies reveal in this case mainly the presence of
amorphous particles in the initial, mechanically activated
powder, although some small crystallites are also observed.
The sizes of these crystalline particles are around 100 nm or
less, like the particle marked with an arrow in Fig. 7. The
inset shows the corresponding electron di!raction pattern,
which only proves the crystalline character of the particle.
Like in the previous cases, the small size together with the
small quantities found made these crystalline particles unde-
tectable by normal XRD analysis. By EDS analysis we
identify crystals containing only Bi or Mo, i.e., coming from
FIG. 6. SEM micrographs of particles from the Bi O :MoO mixture:
(a) starting oxides, (b) mechanochemical activated mixture, (b) after anneal-
ing at 3003C.
ꢀ
ꢂ
ꢂ

60
RICOTE ET AL.
the starting oxides, and others with Bi:Mo ratios almost state reactions, and the absence of other phases usually
2:1, corresponding to Bi MoO , like the one marked in involved in those reactions, have been reported recently in
ꢀ
ꢃ
Fig. 7. The amorphous particles surrounding it are around other systems (28). According to previous results, it seems
500 nm. Their analysis by EDS reveals Bi:Mo ratios close to that the natural trend of mechanoactivated oxides of the
the 2:1 of the "nal composition Bi MoO .
Therefore, it seems that mechanochemically activated synthesized #uorite phases (25), like the Bi VO and
powder in this case contains mainly amorphous particles Bi MoO #uorites in this study. The reasons behind this
with the "nal composition, with traces of crystals of the tendency to obtain new phases by mechanical activation are
Aurivillius family is the formation of new, non-previously
ꢀ
ꢃ
ꢀ
ꢄ
ꢀ
ꢃ
starting oxides and of Bi MoO . Like for Bi VO , we can still not clear, and may be related to the not very well known
ꢀ
ꢃ
ꢀ
ꢄꢅꢄ
conclude that there is evidence of
a
process of mechanisms of mechanical activation, which are di!erent
mechanosynthesis for the system Bi O :MoO .
from the ones ruling traditional thermal activation. For this
ꢀ ꢂ
ꢂ
An annealing process will mainly produce the crystalliza- discussion we assume that in general the formation of
tion of the large amorphous particles. As stated in the case #uorite structures is the most energetically favorable option
of Bi VO , the smaller crystal size observed in the an- for the amorphous, mechanoactivated precursor oxides. In
ꢀ
ꢄꢅꢄ
nealed powder compared to the amorphous particles of the that case, we should consider that the formation of #uorite
milled one can be explained due to shrinkage produced phases requires a change in the coordination of the cation B,
during crystallization from the amorphous particles. This is from the usual 6 of the starting oxide to 8 of the #uorite.
ꢃ>
ꢆ>
also re#ected in previous DTA results (25), where a narrow This is possible for cations like Mo and V , but not for
ꢄ>
peak for the formation of this #uorite phase con"rms this
conclusion.
V
, which has only been reported to have coordinations
5 or 6 (29). Therefore, in the case of Bi VO , it seems
reasonable that mechanoactivated particles react to form
ꢀ
ꢄꢅꢄ
ꢄ>
4. DISCUSSION
a structure with 5 or 6 coordination for V , like
c Bi VO , and not a #uorite.
ꢀ
ꢄꢅꢄ
The activation process in a vibrating mill used in this
The annealing of the mechanoactivated particles at mod-
work, more energetic than the one reported by other erate temperatures (between 250 and 4003C) results either in
authors (23), produces mainly amorphous particles of very the formation of the "nal mixed oxide (f-Bi VO ) if
ꢀ
ꢄ
small sizes ((1 lm) as shown in the SEM micrographs of mechanosynthesis has not taken place, or in the crystalliza-
the n"1 Aurivillius oxides studied here. This is the conse- tion of the already synthesized phase (c-Bi VO f-
ꢀ
ꢄꢅꢄꢇ
quence of both the rupture of the starting particles and the Bi MoO ). The evolution is di!erent in both cases and it is
ꢀ
ꢃ
massive creation of defects by intense milling, which results re#ected in the DTA studies of these compounds (24, 25):
in amorphization, as it has been observed in other systems a narrow exothermic peak corresponds to the crystalliza-
(27). Regardless of the similar composition and structure of tion of c-Bi VO and f-Bi MoO , while wide peaks indi-
ꢀ
ꢄꢅꢄ
ꢀ
ꢃ
the materials studied, di!erent results are induced by the cate the synthesis and crystallization of the f-Bi VO . When
ꢀ
ꢄ
mechanochemical activation used. We "nd evidences of only a crystallization process is involved, shrinkage of the
mechanosynthesis, i.e., synthesis induced only by mechan- particles is observed (Figs. 2c and 6c). If the process involves
ical activation of the starting oxides, for two of the systems the synthesis of a new phase, f-Bi VO , amorphous particles
ꢀ
ꢄ
studied: 2Bi O :V O and 2Bi O :MoO . In both systems, of the starting oxides agglomerate to start interdi!usion
ꢀ ꢂ ꢀ ꢄ
ꢀ ꢂ
ꢂ
the energy introduced in the system is able not only to processes (Fig. 4c). This process of synthesis and crystalliza-
produce the synthesis, but also to nucleate some very small tion seems not to be completed at a de"nite temperature
crystallites. In the third system, Bi O :VO , particles of (2753C in this case). Instead, further small transformations
VO remain crystalline after the mechanical activation pro- take place when more energy is entered in the system
cess, which is most probably the reason for the absence of through a thermal process or by the electron beam.
ꢀ ꢂ
ꢀ
ꢀ
mechanosynthesis in this case. Considering that amorphous
The phases obtained by mechanical activation and de-
particles are highly reactive, if amorphization of one of the scribed above, either have not been synthesized by the
starting oxides has not been fully achieved, the synthesis of traditional methods of thermal activation, or show new
the "nal product will require in principle more energy input. features like the stabilization at room temperature of the
Thus, the chemical reaction between the starting oxides is c phase of Bi VO . Further heating of the mechanoac-
ꢀ
ꢄꢅꢄ
achieved only by annealing. To conclude, mechanosynthesis tivated material takes it back to the phases obtained by the
is possible only when amorphization of all the starting traditional solid-state reaction. For example, c Bi VO
oxides is complete, as this process implies the rupture of remains stable at room temperature until we anneal it above
ꢀ
ꢄꢅꢄ
their initial structure and, consequently, an increase of their 6503C, temperature of formation of this disordered phase
reactivity.
from the ordered b phase in a thermal process. Once we
The formation of compounds by mechanochemical ac- reach that temperature, the c phase disappears on cooling,
tivation, which cannot be obtained by conventional solid- transforming successively to the ordered b and a phases (24),

STUDY OF THE PROCESS OF MECHANOCHEMICAL ACTIVATION
61
the same evolution observed in the material obtained by the
traditional solid-state reaction. Similarly, #uorites Bi VO
ACKNOWLEDGMENTS
ꢀ
ꢄ
and Bi MoO obtained from mechanical activated powders
The authors thank Ms. P. Begue for the synthesis of Bi O :MoO
ꢀ
ꢃ
ꢀ
ꢂ
ꢂ
products. This work has been funded by the Spanish CICYT (MAT97-
0711) and CAM (07N/0061/1998). TEM work has been carried out at the
Centro de Microscopma Electronica &&Luis Bru'' (Universidad Complutense
de Madrid, Spain).
transform on heating into the same Aurivillius phases ob-
tained when the material is produced by thermal activation
(24, 25). These results evidence not only that the mechanisms
of phase formation are di!erent for mechanical and for
thermal activation, but also that the structures obtained by
mechanoactivation can be reversed to the ones obtained by
traditional thermal activation. We have shown that the
combination of mechanical and thermal activation allows
the production of both new and previously reported phases
of the systems studied here.
REFERENCES
1. B. Aurivillius, Ark. Kemi. 1, 463}480 (1949).
2. A. F. van den Elzen and G. D. Rieck, Acta Crystallogr. B 29, 2436}2438
(1973).
3. A. A. Bush and Yu. N. Vanetsev, Russ. J. Inorg. Chem. 31(5), 769}771
(1986).
4. V. G. Osipyan, L. M. Savchenko, V. L. Elbakyan, and P. B. Avakyan,
Izv. Akad. Nauk SSSR Neorg. Mat. 23(3), 523}525 (1987).
5. I. H. Ismailzade, I. M. Aliyev, R. M. Ismailov, A. I. Alekberov, and
D. A. Rzayev, Ferroelectrics 22, 853}854 (1979).
6. D. A. Je!erson, J. M. Thomas, M. K. Uppal, and R. K. Grasselli,
J. Chem. Soc. Chem. Commun. 594}595 (1983).
7. K. Shantha and K. B. R. Varma, Solid State Ionics 99(3}4), 225}231 (1997).
8. P. B. Avakyan, M. D. Neresyan, and A. G. Merzhanov, Am. Ceram.
Soc. Bull. 75(2), 50}55 (1996).
9. K. Shantha and K. B. R. Varma, J. Mater. Res. 14(2), 476}486 (1999).
10. P. Millan, J. M. Rojo, and A. Castro, Bol. Soc. Esp. Cera& m. <idrio 37,
217}220 (1998).
11. P. Millan, J. M. Rojo, and A. Castro, Mater. Res. Bull. 35(6), 835}845
(2000).
5. CONCLUSIONS
An analysis of several mechanochemically activated n"1
Aurivillius oxides in a vibrating mill was carried out by
SEM, TEM, electron di!raction, and EDS. The mechanical
activation results in the formation of submicrometer
amorphous particles by the rupture and massive creation of
defects of the precursor particles. However, di!erent results
are induced by the mechanical activation, regardless of the
similar compositions and structure of the compositions
studied.
In the systems 2Bi O :V O and Bi O :MoO , evidence 12. F. Abraham, M. F. Debreuille-Gresse, G. Mairesse, and G. Nowog-
ꢀ ꢂ ꢀ ꢄ
ꢀ ꢂ
ꢂ
rocki, Solid State Ionics 28+30, 529}532 (1988).
13. S. Sorokina, R. Enjalbert, P. Baules, A. Castro, and J. Galy, J. Solid
State Chem. 125, 54}62 (1996).
14. P. Shuk, H. D. Wiemhofer, U. Guth, W. Gopel, and M. Greenblatt,
Solid State Ionics 89, 179}196 (1996).
of mechanosynthesis taking place as a result of the activa-
tion procedure was found. The energy introduced in the
system not only produces amorphous particles containing
all the elements, but also small crystallites of the "nal phases
are identi"ed. Further annealing of the product results in 15. M. E. Arroyo y de Dompablo, F. Garcma-Alvarado, and E. Moran,
Solid State Ionics 91, 273}278 (1996).
16. G. Mairesse, in &&Fast Ion Transport in Solids'' (B. Scrosati, Ed.),
p. 271. Kluwer Academic, Dordrecht, 1993.
17. K. R. Kendall, C. Navas, J. K. Thomas, and H. C. zur Loye, Chem.
Mater. 8, 642}649 (1996).
the complete crystallization of the particles. However, a sim-
ilarly processed powder of Bi O :VO produces syn-
ꢀ ꢂ
ꢀ
thesized particles only when annealed. Particles of VO do
ꢀ
not amorphize after the prolonged milling, the rupture of
the initial structure does not take place, and, as a conse- 18. K. Tkacova, &&Mechanical Activation of Minerals.'' Elsevier, Amster-
dam, 1989.
quence, mechanosynthesis is inhibited. Therefore, we con-
clude that mechanosynthesis is possible only when
amorphization of all the starting oxides is achieved.
Mechanochemical activation of these compounds pro-
19. V. V. Boldyrev, Solid State Ionics 63+65, 537}543 (1993).
20. J. J. Gilman, Science 274, 65 (1996).
21. J. M. Gonzalez-Calbet, J. Alonso., E. Herrero, and M. Vallet-Regm,
Solid State Ionics 101+103, 119}123 (1997).
duces either non-previously synthesized #uorites of Bi VO
22. P. Lacorre and R. Retoux, J. Solid State Chem. 132, 443}446 (1997).
23. K. Shantha and K. B. R. Varma, Mater. Sci. Eng. B 60, 66}75 (1999).
24. A. Castro, P. Millan, J. Ricote, and L. Pardo, J. Mater. Chem. 10,
767}771 (2000).
25. P. Begue, P. Millan, and A. Castro, Bol. Soc. Esp. Cera& m. <idrio 38(6),
558}562 (1999).
ꢀ
ꢄ
and Bi MoO , or a previously reported Aurivillius phase
ꢀ
ꢃ
ꢄ>
when one of the ions, V , cannot "t into any #uorite
structure. Even in this case, there is a novelty, the c-
Bi VO
ꢄꢅꢄ
phase is stable at room temperature. All this
demonstrates that the mechanisms of phase formation are 26. A. Castro, P. Millan, L. Pardo, and B. Jimenez, J. Mater. Chem. 9,
ꢀ
1313}1317 (1999).
di!erent for mechanical and for thermal activation. Further
annealing takes all these structures back to the ones ob-
tained by traditional thermal activation, which reveals how
27. G. J. Fan, F. Q. Guo, Z. Q. Hu, M. X. Quan, and K. Lu, Phys. Rev.
B 55(17), 11010}11013 (1997).
28. J. Wang, J. Xue, and D. Wan, Solid State Ionics 127, 169}175 (2000).
the structures obtained by mechanoactivation can be rever-
sed to the ones obtained in thermal cycles.
29. R. D. Shannon and C. T. Prewitt, Acta Crystallogr. B 25, 925}946
(1969).